BIOMIMETIC TRANSFER OF PLANT ROOTS FOR PLANETARYANCHORING

Tobias Seidl1,∗, Sergio Mugnai2, Paolo Corradi3, Alessio Mondini3, Virgilio Mattoli3, ElisaAzzarello2, Elisa Masi2, Camilla Pandolfi2, Barbara Mazzolai3, Cecilia Laschi3, Paolo

Dario3, Stefano Mancuso2

1) European Space Agency, Advanced Concepts Team, ESTEC, Keplerlaan 1, Postbus 299, 2200 AG,Noordwijk, The Netherlands; 2) LINV, Dept. Horticulture, University of Florence, viale delle Idee 30,50019 Sesto Fiorentino (FI), Italy; 3) Scuola Superiore Sant’Anna, V.le R. Piaggio 34, 56025 Pontedera

(PI), Italy

∗ Corresponding author, Email: tobias.seidl@esa.int

Abstract

The task of autonomously anchoring spacecraft on the surface of any celestial body is extremelychallenging. Weight restrictions do not allow for massive design; lacking external control calls for com-pliant drilling behavior. In search for a new concept we investigated biological mechanisms of groundanchoring.Plants are the first organisms to settle in any type of empty habitat. The roots play an importantrole during settling providing scalable anchoring. This feature is a consequence of longitudinal growthprocesses at the tip of the roots (apices) and radial growth on the overall root allowing them to increasein size and strength as well as adapting to changes in external loads. The growth direction of the apexis controlled by a simple set of mechanical and chemical sensors in the transition zone. Root apicesalso exchange information between each other for coordination.We transfered the technical features of roots concerning anchoring strategy, actuation, and controlaspects into engineered concepts. Extending the ’root’ at the tip reduces friction with the substrate,a new type of osmotic actuation works without few moving parts and consumes low energy. A dis-tributed control architecture allows for individual and intra-individual steering, negotiating obstaclesand ensuring tight anchoring. Finally we present a technical root integrating the current technologicalpossible solutions.

INTRODUCTION

On first sight it appears strange to focus on livingnature in search of new design concepts for space-craft. However, plants represent a major fractionof living beings on Earth and have conquered al-most any surface on our planet. Although theycannot actively move, plants are the first settlersin a hostile environment, making path for a habi-tat that can then be settled by all kinds of ani-mals. After having been placed into a new envi-ronment - usually as a seed and being transportedpassively - they need to anchor in the substrateand exploit the available resources in order to grow.From this description it becomes clear, that rootsplay an important role for a successful establish-ment on a new site. The roots need to provide me-chanical anchoring, preferably adaptive to chang-ing size of the stem above ground. Each single roothas to move through the substrate, orienting alongthe gravity vector, negotiating obstacles, and lo-

cating resources at the same time while balancingthe external loads applied. The entire behavior isachieved by an osmotic actuation system that issteered by a distributed set of simple controllers lo-cated in the tip of each root - the apex. During aplanetary mission, a spacecraft can be seen as goingthrough similar stages as a plant in the beginningof its life. After having landed, the once mobilespacecraft might anchor and probe the substratefor scientific reasons. Inspired by this analogy, ateam of biologists and engineers investigated boththe actuation and the control mechanisms of plantroots in the focus of a biomimetic transfer for con-ceptually novel anchoring solutions for exploratoryspacecraft [1].

Plants and their roots

Plants usually have an aerial part and a root sys-tem. Plant roots of, e.g., the mesquite (genusProsopis) may extend down more than 50 m to

1

reach groundwater. Annual crop plants develop aroot system that can usually grow until 2.0 m indepth and extend laterally to distances up to 1.0 m.As a general rule, the annual production of rootsmay easily surpass that of shoots, so the above-ground portions of a plant can be seen as the tipof an iceberg. As plants are mostly sessile, theyhave developed certain growth responses - so calledtropisms - which allow them to respond to changesin their environment. The best known stimuli fortropisms are light, gravity, touch, and humidity[2, 3]. Together, these rather simple mechanismsallow plants to conquer almost each type of environ-ment. In doing so, root systems often far exceed inmass and length the above-ground portions of theplant, being provided with few to many main rootsand thousands of ramifications. In spite of theirdelicate structure, the spiraling forward thrust ofthe root tips and the pressure of their expandingcells are sufficient to split solid rock. Plant tissueexpands in tight places by taking up water via os-motic mechanisms, leading to an overall increase insize both by cell expansion and division. In order toprovide firm anchoring, the directions in which theindividual root ramifications grow need to be coor-dinated, let alone the individual avoidance of ob-stacles like, e.g., stones. Since an established rootcannot change its shape any more, growth directioncan only be achieved at the tip of the root, the apex.Currently, there is a general agreement that higherplants are not only able to receive diverse signalsfrom the environment but that they also possessmechanisms for rapid signal transmission [4]. Infact, each root tip is able to receive informationfrom the environment via embedded sensors; the in-formation is then transduced and processed in thewhole root system and used to direct the growth to-ward regions of the soil with, e.g., the best mineralsand water availability. So plants can effectively pro-cess information obtained from their surroundingsand can show a learning behavior which involvesgoal seeking, error-assessment, and memory mech-anisms [5, 6, 7]. Therefore plants demonstrate tosuccessfully reach their needs even without a con-ventional locomotion system. In consequence, rootapices are not only sites of nutrient uptake but alsosites of forward movement. During growth, plantroots may exert pressures of up to 1 MPa in orderto penetrate hard soils [8]. Consequently, virtu-ally all plants which grow in soil have evolved root

caps which protect the root meristem from phys-ical damage or abrasion by soil particles. Thesecaps play an important role not only in protectingthe root meristem from damage, but also in deter-mining the mechanical interaction between the rootand the soil, namely the mode of soil deformationand the rootsoil friction [9].

Scenario for a biomimetic plant

In the present account we focus on planetary an-choring solely without taking into consideration is-sues of an entire space mission such as launch re-quirements, space navigation, etc. It may well beassumed that the earlier stages of our proposedscenarios would be qualitatively similar to thosepresented to current planetary probes. However,we consider a significant change during surface ap-proach including a rather hard - and obviously morerisky - landing which would provide the probe withan initial, dynamic anchoring in the substrate simi-lar to that of common plant seeds. Any soft landingwould require an additional fixation of the probeto facilitate first penetration of the substrate bythe extending roots. As we will see later, an au-tonomous penetration of the ground will require acloser imitation of plant roots than currently pos-sible. A plant-robot, or eventually many plantoids,disseminated like seeds in large lands, extendingtheir roots in the ground, could autonomously anal-yse the composition of soil and monitor the pres-ence of a variety of chemical-physical parameterseven in sub-surface locations. The main features ofthe plant-bot for sub-surface planetary exploration,are minimal power consumption and a rather com-pliant and friction-less penetration compared torigid drillers which would preserve fragile samplesto a better extent.

The characteristics of the substrate to be ana-lyzed are important for the success of the pene-trating principle of the roots. It is likely to belimited to a loose and fine sand and soil, whilethe plantoid will intuitively not able to perforatehard bedrock. Indeed, the plantoid should be ableto detect and avoid such obstacles. We identifiedthree primary space mission targets: Mars, Moon,and Asteroids. Among these targets, Mars explo-ration appears to be the easiest to realize, thanksto the less strict constraints. On Mars tempera-tures range from -87C to -5C [10], and the surface

2

level atmospheric pressure at the surface level av-erages around 0.6 kPa (i.e. less than 0.01 timesthe terrestrial atmosphere). The top layer of mar-tian soil has an estimated thickness of 0.5 to 1 mand is expected to be composed of fine sand andloose soil. When reaching deeper more hard rocklayers can be expected carrying stones correspond-ing to those of our mafurite, carbonatite, diopsideand diabase [11, 12, 13]. In consequence, the rootsneed to be sufficiently versatile and flexible to avoidstones and hard bedrock. Even if the first layer onMars consists in an iron oxide dust with the con-sistency of talcum powder there might be water iceclose to the surface as was indicated by the MarsOdyssey orbiter [14] and recent findings from theUS Phoenix mission [15]. Although this ice is notexpected to form a rigid layer but rather be presentin a grainy condition we may have to consider ad-ditional measures to facilitate penetration.

Our goals were (i) to achieve an integrative un-derstanding of the technical aspects of plant rootsas far they were relevant for our project followed by(ii) a biomimetic transfer into feasible technologicalconcepts as well as (iii) developing a technologicalroad-map for the ’ideal’ biomimetic root. The twomain aspects of this study were (i) the growth ofplant roots and (ii) the control issues involved insteering of the growth directions. The cell expan-sion is achieved via rather slow osmotic processeswhich require little energy and are actuated for-ward with rather low friction. The several differentindividual roots of one plant need to individuallyfind their way and coordinate with their neighbors.Both control within and communication betweenroot apices are examined for a bioinspired controlarchitecture. In the present account we report onthe results of our joint work in transferring plantknow how into technological concepts.

ANCHORING CAPABILITY

State of the art

Anchoring is becoming a critical aspect of plan-etary missions involved with in-situ explorationtasks, particularly on planets with a very low grav-ity, where the low weight force of the spacecraftcould not be sufficient to counter-balance the re-action force generated by the terrain due to the

soil probing, which could exert a significant axialpush. However, anchoring systems are not well rep-resented within the scientific literature available inthe space sector. Mostly they consist of metallicthorns or nail like structures that extrude rigidlyfrom the landers landing pads and are rammed intothe ground during landing. Relevant examples ofthose can be found in [16] and [17]. All rigid an-chor systems face the problem of performance overa longer time frame. If the soil creeps, reliable an-choring will not be granted any more. Also the sam-pling and probing techniques will face similar is-sues. Natural anchoring and sampling as exhibitedby growing plants have the ability of continuousadaptation by longitudinal and radial extension, in-cluding changing of growth direction. A technicaltransfer of root growth may seem visionary but isdesirable in terms of improved performance.

Anchoring issues

As a primary rule, the axial force generated by theanchoring process onto the terrain must be evi-dently lower than the weight of the spacecraft toprevent lifting it. As introduced before, the an-choring system could take advantage of the entryphase of the probe by exploiting a preliminary self-insertion into the terrain of the lower parts of thespacecraft due to impact landing on the planetarysurface.This should guarantee a pre-anchoring inorder to allow the extension of the robotic root andthus achieve progressively a more stable anchoringcondition while eventually probing the soil for sci-entific investigation.

Relying on the described osmotic actuation prin-ciple, the apex would be hopefully able to steeronly at the very beginning of the sandy-soil surface.Going deeper, the steering momentum requested inorder to move the surrounding sand could be toolarge. Hence, the steering actuation should occurdynamically, that means progressively during thequasi-static forward penetration of the root. Theresistance itself that the terrain will oppose to theroot penetration will contribute to steer the apexinto the wanted direction, producing the change inthe desired direction of the penetration. This ap-proach might actually lead to a condition whereeventually only a single bending of the robotic rootcould occur, insufficient for a suitable soil probing,but probably enough for an anchoring purpose. In

3

order to evaluate the steering capability and thementioned issues, tests in soil are definitely neces-sary.

Considering to push the entire root from the sur-face (as it happens in the Cone Penetration Tests,CPTs, in the Industrial field, where straight probesare axially pushed into the terrain for several me-ters), it must be observed that the penetration of abended root would be actually even more difficult,because the pushing force will not be entirely prop-agated down to the apex, but only a component ofthe force will contribute to move the apex forward.In order to compensate this problem, the root pene-tration can be conceived by an initial straight inser-tion of several root modules into the terrain whichis followed by a serial actuation of the undergroundmodules from the upper stages down to the apex,in order to avoid to push all the already extendedmodules; this should decrease the problem of ad-vancing all the root through a bended path (thatmeans also decreasing the sleeve friction while theroot is penetrating).

It must also be considered that on Mars, for in-stance, the gravity is about one third the gravityon Earth and the mean surface level atmosphericpressure is less than 0.01 the Earth pressure. Thisshould be a more favorable condition to the soilpenetration compared to Earth tests. On the otherhand, this is a unfavorable condition regarding ini-tial anchoring capability. Experimental trials insoils could actually simulate such a condition anddemonstrate if it exists a suitable compromise be-tween force requested for the penetration (which inturn relies on the root size considering also the ex-ploitation of the osmotic actuation principle) andanchoring capability in order to prevent lifting ofthe system with a specified weight.

ACTUATION

Mechanistic understanding of root growth needsto take into account the architecture of the rootgrowth zone. As a given region of the plant axismoves away from the apex, its growth velocity in-creases (the rate of elongation accelerates) until aconstant limiting velocity is reached equal to theoverall organ extension rate. In a rapidly growingmaize root, a tissue element takes about 8 hours tomove from 2 mm (the end of the meristematic zone)

to 12 mm (the end of the elongation zone), with amain velocity of 3 mm/h. In physical terms, cellgrowth can be defined as an irreversible increase incell volume and surface area. Plant protoplasts arecharacterized by an univocal structure and organi-zation, as they are surrounded and encased by arigid, but expandable, cell wall with elastic charac-teristics. The cell wall is infiltrated with water con-taining only a very low osmotically active amountof solutes. This situation enables the formation of alarge difference in osmotic pressure (∆Π) betweenthe apoplastic (outer) and the symplastic (inner)space (0.6-1.0 MPa), compensated by a hydrostaticpressure, named turgor (P ), of equal value whenthe cell is in a fully turgid state. The cell canhence be described as a simple osmometer that canenlarge by water uptake powered by a differencein water potential (∆Ψ) between protoplastic andsymplastic space.

Growth in plant roots is a result of two mecha-nisms: (i) cell division in the apical meristem justbehind the tip, and (ii) cell elongation in a zone justbehind the apex [18]. The driving force for cell elon-gation is an increased turgor (i.e. pressure) throughwater influx into the cell (Figure 1). This waterinflux is a result of the osmotic potential withinthe plant cell located in the elongation zone [19].The water relations of expanding cells have beenreviewed in detail [20]. A typical value for the vac-uolar osmotic potential (equal to Πi) inside a cellin the growing zone of an unimpeded root grownin hydroponics is around -0.7 MPa (about 7 Atm).Classically, following experiments on cell walls iso-lated from giant algal cells [21], cell elongation hasbeen regarded as plastic flow of the wall materialunder stress [22]. Where existing soil channels aresmaller than the root diameter, roots must exerta growth pressure in order to displace soil parti-cles, overcome friction and elongate through thesoil. The growth pressure σ is equal in magnitudeto the soil pressure that opposes root growth. Ina root tip elongating through soil, cell turgor pres-sure P generates the growth pressure σ, which re-sults from the difference between P and the wallpressure W [23].

In unimpeded roots, σ is by definition zero andP is balanced by W . When roots are completelyimpeded and cannot elongate, σ attains a maxi-mum value σmax. Roots exert growth pressures inboth radial and axial directions, but we will only

4

Figure 1: Schematic representation of a root grow-ing in soil. The turgor acts against soil resistance.

deal with axial growth pressures here. When wa-ter transport into growing cells is not limiting, rootelongation rate can be considered in terms of a sim-plified Lockhart equation [22], as modified by [23]to take account of the soil impedance:

R = m[W −Wc] = m[P −Wc− σ]

where R is the elongation rate, m is a cell wall ex-tensibility coefficient, W is the wall pressure, Wc isthe cell wall yield threshold, P is the turgor pres-sure and σ is the soil impedance (or growth pres-sure). As shown above, the mechanism of root cellexpansion is driven by the osmotic pressure. Thisosmotic expansion mechanism could be successfullyused to design a new class of electrochemical actu-ators able to fit the necessary requirements (lowpower, slow actuation, high force/pressure). Theconcept of this new electro-osmotic actuators in-spired by plants will be presented in the next sec-tions.

The theoretical working principle of the plantinspired electro-osmotic actuator is now describedin detail. Figure 2 shows two cells connectedthrough a semi-permeable membrane. Each cell hasa metallic electrode (M and N) immersed in a wa-ter solution of own ions (M+ and N+, respectivelymolal concentration cM+ and cN+ ) in presence of acounter-ion (X−), required for the electroneutral-ity of the solution. If the concentration of the twocells is different, an osmotic pressure is generatedacross the semi-permeable membrane. In this casethe osmotic pressure Π is given by the Van tHoffslaw [24]:

Π = Φ (cM+ − cN+)R · T

where R is the gas constant (0.082 l · atm ·mol−1),T the temperature (in Kelvin) and σ is the molalosmotic coefficient, a coefficient that takes into ac-count the non-ideality of the solution. On the baseof the equation (2) it is possible to evaluate the the-oretical pressure achievable by using the osmoticmechanism. Considering a difference of electrolyticconcentrations between the two cells of 1M (a real-istic achievable concentration) and considering themolal osmotic coefficient σ = 1, we obtain a pres-sure of about 2.5MPa (24 Atm). The force thatcan be exploited by this pressure is obviously es-sentially related to the geometry and dimensions ofthe transduction mechanism. Few examples of de-vices that use pure osmotic actuation mechanismare available in literature [25].

To control the displacements and the forces gen-erated by such kind of osmotic actuator it is neces-sary to control the ion concentrations on the cells.This goal can be achieved by using electrochemicalreaction. If the membrane that connects the twocells allows the passage of some kind of ions (notsimply a semi-permeable membrane), thus the cellscan work as a battery (it can generate an electri-cal current) or conversely work as electrolytic cells(applying a suitable external voltage).

The two reduction semi-reactions involved in theprocess are:

M+ + e− → M

N+ + e− → N

The electric potential across the two electrodes ∆Eis given by the Nernst equation:

∆E = EM − EN

=(E0

M − E0N

)+

R · TF

ln ([cM+ ] / [cN+ ])

where E0M and E0

N are the semi-reaction standardred-ox potentials, and F is the Faraday constant(F = 96, 485.34 s · A · mol−1). Providing a suit-able current to the cells it is possible to control theconcentrations of the electrolytes. The electricalpotential needed to induce a current in the cells isin the order of few Volts (at considered concentra-tions). The induced variation of electrolyte concen-tration within a cell, after application of a currentI for a time t, is given by the following equation:

cM+(t)− cM+(0) =I · t

F · V

5

Figure 2: Working principle of a electro-osmotic actuator.

where V is the volume of the cell. It is important tounderline that in front of the variation of concentra-tion of the active electrolytic chemical species (inthis case the cations M+ and +) due to the red-oxprocess, the ionic transport of counter-ions (in thiscase X−) must be allowed (otherwise electrode getspolarized), by means of a ion selective membrane.Clearly the active ionic species must be blocked bythis membrane otherwise a net variation of concen-tration is not achievable.

Since the underlying physico-chemical processesare rather well understood we were able to identifycandidate component for a hard-ware realisationof an electro-osmotic actuator: Two chemicals -Pb(ClO4)2 or in alternative PbSO4 - can be used toprovide the relevant ions to drive the process. How-ever, we were not able to identify a membrane thatwould combine ion-selectivity and osmotic perme-ability at the same time. Hence, composite mem-branes had to be evaluated. In this context, weperformed some very preliminary experiments tomonitor the performances. In a first set of experi-ments performed on these membranes we have ob-tained that with 2 chambers of 15 x 60 x 60 mm, 1

molar solution of NaCl in H2O and 12.5 mm2 ex-change surface, the highest pressure has been ob-tained with Cellulose Acetate RO CE membrane,11 Atm reached in 3 hours. This pressure is com-parable with the growing pressure generated by theplant roots. Similar test will be performed withlead in order to evaluate the maximum pressuresobtainable with a controlled process. In combina-tion with anionic exchange membranes we achieveda satisfactory selectivity response and relatively lowresistance (allowing reasonable ionic flows), makingthem useful for our application.

CONTROL

The root system, which is the plant organ devotedto soil exploration, is composed by thousands ofroot tips or apices. Each apex receives informationfrom the environment via embedded sensors anduses it to direct growth toward soil regions withhigh mineral and water availability. The transitionzone of the root apex - situated between divisionand elongation zone - is able to detect more than 10

6

chemical and physical environmental parameters.In this zone, the cells exhibit a unique cytoarchitec-ture with centralised nuclei surrounded by perinu-clear microtubules radiating toward the cell periph-ery. This particular configuration should be suitedfor the perception of external stimuli and a sub-sequent transmission to the nuclei. The transitionzone is also regarded as the location of informationprocessing, decision taking, and information stor-age due to its high energy consumption in form ofATP.

The growth of root tissue relies on the establish-ment of cellular and sub-cellular asymmetries. Theformation of the real cell shape and the positioningof molecules in the intracellular space commonly in-volve a persistent directional orientation along anaxis, named cell polarity. Polar transport of auxinis directly linked as a signal to the regulation ofboth growth and polarity in the plants. As theplant body is shaped in response to numerous en-vironmental stimuli (i.e. light and gravity; [26, 27],these factors are able to influence the transport ofauxin in a way that this hormone is delivered totissues induced to grow via the establishment ofauxin gradients, transport and response. Auxin istransported across the whole plant body via effec-tive cell-cell transport mechanisms involving boththe symplast and the apoplast. Cellular auxin in-flux and efflux, and the mechanisms that mediatethe delivery and removal of potential polar auxintransport components from the plasma membrane,remain still open and discussed. For example, itis not clear why auxin bypasses the cytoplasmaticchannels of the plasmodesmata crossing throughthe apoplast, as their diameter could easily ac-commodate several auxin molecules. This suggeststhe presence of an active mechanism that preventsauxin entering the plasmodesmata [28] and im-plies a functional benefit for including an apoplasticstep in the polar transport of auxin. Transcellularauxin transport is accomplished via a poorly un-derstood vesicle-based process that involves the pu-tative auxin transporters, or transport facilitators,recycling between the plasma membrane and theendosomes. Active auxin transport mediates cellu-lar auxin concentration and is therefore a crucialcomponent in the coordination of plant develop-ment. However, the specific relationship betweenauxin signaling and auxin transport is still quiteunknown.

Plants, along their evolution, have had to faceand solve a variety of problems to survive: de-veloping support structures, creating a system oftransport of water and nutrients through the wholeorganism, protecting the delicate reproductive or-gans, assuring the reproduction in the most con-venient period of the year, improving the mecha-nisms of adaptation to the variable climatic condi-tions etc. All these needs has led to large modifi-cations in metabolism and has brought to the ac-quisition of sensorial structures to set up a precisebiological clock necessary to guarantee the constantmonitoring of the surrounding situation to acclima-tize them. Plants can sense gravity, temperature,light quality and direction, etc., and if necessarythey can act consequently. Plants feed themselves,breathe, fight the infections, in some cases gener-ate symbiosis with fungi and bacteria and commu-nicate with them. They live in continuous com-petition both with environmental agents and withother plants for the conquest of light, space andnourishing substances, and all those necessary ele-ments to their survival. They also live in contin-uous competition with other predator organisms,fungi, bacteria or animals. When circumstances be-come unfavorable for optimal growth and develop-ment of animals, they can respond accordingly bymoving to a more favorable environment. Plantsare not afforded this luxury. Due to their sessilenature, plants are forced to make the most of theirimmediate surroundings, which means adapting toan ever-changing environment [29]. So they havebeen adapted to perceive and react to adverse sit-uations with varies useful movements to the sur-vival. Darwin noted that plants had a tendency tosense their environment so as to orient themselvesfor optimal growth and development, and he dedi-cated part of his studies to vegetal biology publish-ing with his son Francis a fascinating book ”ThePower of Movement in Plant” [30] in which he en-closes many interesting observations on plant lifewith special interest in movement. The Darwinsstudied the two great categories of ”movements inthe plants”: ”tropisms (directional movements inanswer to external directional stimuli) and ”nas-tic movements” (movements in answer to externalstimuli, but independent to their direction). Plantsare constantly being bombarded with changes intheir environment. Temperature fluctuations, notenough light or water content in the soil, are just a

7

few of the factors to which plants must be able torespond. Moreover, plants must respond to physi-cal forces of nature such as gravity or touch stimula-tion. Over evolutionary time, plants have adaptedto their surroundings with a high degree of plastic-ity, affording them the ability to respond to ever-changing conditions that provide constant stimu-lation. Plant tropisms are operationally defined asdifferential growth responses that reorient plant or-gans in response to direction of physical stimuli. Anexample of tropism regards the perception of grav-ity, also called Gravitropism. Phototropism is thedirectional response in answer to a luminous source.Tropisms can be negative, such as a stem bendingaway from a gravity stimulation [31], or they can bepositive, as in a stem bending toward a light stim-ulation [29] while example of nastic movements arethe closing leaves, modified as fly-trap, of Dioneamuscipola, the fast closing of the composed leaves ofMimosa pudica to a tactile stimulus, or the closingleaves of many leguminous during night (nictinas-tic movements). Darwin studied also another typeof movement, oscillating and rhythmic, called ”cir-cumnutation”: plants, during their growth, per-form circular movements around a central axis [32].Nearly all the plants present this growing move-ment, however it turns out to be more obvious insome species rather than in others, such as climb-ing plants [33, 34]. Therefore tropisms are thatgroup of reaction mechanisms that the plant actsin relation to directional stimuli. The most knownand studied are: phototropism the answer to light;gravitropism the answer to gravity; electrotropismthe answer to an electric field, hydrotropism theanswer to a gradient of water in soil.

In the earth gravitational field the roots grow to-ward the bottom for better supplying water andnutrients (gravitropism). As gravity is a physicalforce which only act on masses, several organellesor particles which are denser or lighter than thecytoplasm can be involved in gravity sensing. Thestages of gravitropism in plants can be divided into:perception of the signal, transduction, and response[35]. In roots, gravity perception occurs in the roottip and the response (differential growth) in thezone of elongation, involving the plant growth hor-mone auxin. In higher plants, there is a temporaland spatial separation between perception and re-sponse, and the signal must be transmitted over arelatively large distance. Two models are proposed

for gravity perception. (1) The starch-statolith hy-pothesis as shown in Figure 3 proposes that per-ception is mediated by dense organelles (statoliths,amyloplasts located in the root tip cells). (2) Theprotoplast pressure hypothesis suggests that theentire mass of the cytoplasm participates in per-ception. The mechanism through which the planttransforms the indication of cell position into a bio-chemical message is still under discussion. An im-portant question in sensor physiology is the deter-mination of the threshold dose of a stimulus in or-der to provoke the reaction of the plant organ. Forthe gravireaction of plant roots, it has been noticedthat gravitational stimuli which last 1 s (perceptiontime) can be actively perceived by plant roots, butone single stimulus is not sufficient to determineany organ response. So, a minimum time of con-tinuous stimulus must occur (presentation time).When a root is subjected to a change of orienta-tion in the gravitational field, the stimulus is per-ceived in less than 1 s. However, this 1 s stimulusis not sufficient to induce a gravitropic response itmust be repeated about ten times (or the stimu-lation must last 10 s) to initiate an asymmetricalsignal within the statocytes (transduction phase).This leads to a downward lateral movement of thegrowth hormone auxin (transmission phase) that isthe cause of the differential growth occurring aftera latent time of 10 min. The four phases are rep-resented in sequence but perception, transductionand transmission can persist for the whole periodof the gravitropic reaction.

Plantoid controller

As a mechatronic system, the robot will be built ona modular scheme. The plantoid can be divided intwo major sections: a first part, which correspondsto the trunk and leaves and it is the upper sectionof the robot, located out of the soil; a second part,which represents the roots, able to move in the soil.In particular, the upper part of the plantoid aims at(i) acquiring energy from the sun for the working ofthe robot, (ii) storing the energy, (iii) transmittingthe data concerning the soil analysis to a remotestation, (iv) managing the plant at high level, and(v) eventually storing the fuel (water or other) ofthe osmotic process for root growth.

The energy is collected by solar panel; the sur-face of this panel is quite small (depending on the

8

Figure 3: Functional scheme of mechanosensing inroot statocytes. The statoliths are able to open theStretch-Activated Ca2+ channels (SACs) throughan exerted tension on actin filaments (AF) or apressure on bridging filaments (BF) that link SACs.(Modified from [35])

solar radiation on the planet) because the powerrequired by the osmotic actuators is very limited.In particular, for the root apex the current con-sumption (osmotic actuator + microcontroller andsensors) can be estimated around few mA with asupply voltage of few volts. For the entire root,e.g., composed of 10 modules, it can be roughly es-timated a power consumption on the order of tensof mW. The major consumption comes from thetransmission block; nevertheless, the data to besent are not many (the plant movement is very slowand, consequently, the soil analyzed is limited andthe data to be acquired are few) and they can betransmitted in a unique solution during the hourof major solar irradiance. The power consumptiondepends on the type of transmission. For exam-ple, the power necessary to transmit to a base lo-cated on the planet surface is different from thepower necessary to transmit to an orbiting space-craft. However, these issues are well known in thespace field. The energy collected by the solar pan-els is stored in a battery in order to have energyduring the absence of solar radiation. Moreoverin the trunk of the robot, there are located alsothe actuators for commanding the pull out of the

solar panels and the tank of fuel for the osmoticprocess. These components are managed by a mi-crocontroller, which performs several functions:

• It controls the high level tasks of the plantoid;

• it collects the data coming from the roots andit uses them to indicate to the roots the portionof soil that must be analyzed;

• it manages the transmission module;

• it commands the actuators for the solar panels;

• it manages the osmotic tank in order to indi-cate to the roots when growing.

Figure 4: Elements of the plantoid involved in op-erating the technical roots.

The lower part of the plantoid is located into thesoil and it is composed by the robotic roots. Theplantoid roots will be able to grow following dif-ferent stimuli, such as gravity direction, in orderto explore the environment in terms of presence ofa variety of chemical-physical parameters and lifesignatures. Each root is formed by an apex thatcomprises sensors and the control part electricallyconnected to the main microcontroller in the plantbody, and by an elongation zone that connects me-chanically the apex and the trunk of the plantoid.Each apex embeds a microcontroller module forthe emulation of the roots behavior through thelocal implementation of networks using as mod-els the real apexes behavior. By imitating theplants strategy, the robot will move slowly, explor-ing efficiently the environment and showing highactuation forces and low power consumption. Theplantoid apex will grow and move into the ground

9

Figure 5: Schematic bi-dimensional overview of the technical root-concept at its current state: pen-etration is conceived by the use of osmotic axial piston-like modules that are connected by means ofcontrolled steering stages.

through the new electrochemical actuators, basedon the variation of the osmotic pressure in a liq-uid, controlled by small electrical signal application(electro-osmotic actuators). These actuators allowa root movement on the plant time scale, applyingrelative high forces with low power consumption.The plantoid will be able to expand and actuate itsroots, implementing de facto a biomimetic growingmechanism. All the roots of the plants are thenconnected to the central body to realize a networkthat drives the growing of the roots in preferentialdirections, driven by the information acquired bythe sensors on the apices.

Plantoid sensors

As already described, the roots and in particularthe root apexes reply to the changes in the envi-ronment through mechanisms that are known astropisms. In detail, the apex has sensors for thegravity (statolith), for the soil moisture and chem-icals, which produce respectively gravitropism, hy-drotropism and chemotropism. In order to imitatethe plant behavior, the following components areconsidered (compare Figure 4):

• an accelerometer to replicate the capability ofthe root to follow the gravity;

• a soil moisture sensor to follow the possiblegradient of humidity in the soil;

• a microcontroller to realize the distributedcontrol of the plant (every apex is an indepen-dent unit);

• a number of actuators (osmotic) for the steer-ing and the penetration of the root in to thesoil;

• other sensors to perform chemical analysis ofthe soil.

CONCLUSIONS

The present concept of a technical root - integratedinto a plant-robot called plantoid - mirrors a num-ber of technical features found in plants (Fig. 5).With some initial insertion, the plantoid can extendits roots and drive them slowly into the ground.The energy delivered by the solar modules is suffi-cient to drive the osmotic actuators. The sequentialarrangement of the actuators allows for a reducedfriction when pushing the roots into the soil. Simi-larly actuated joints provide compliance. The con-troller is sufficiently small to be integrated into thetip of the root. At this first attempt to mimic rootsin a technical concept these feature are unique. Sofar we are aware of only one other attempt to mimicroots for planetary anchoring [36]. However, theapproach differs considerably as it integrates the bi-ological technologies far less. In the future the herepresented concept awaits experimental evaluationwhich will definitely lead to further optimization.

But the visual comparison between the thin andnumerous roots of plants and the rather bulky de-sign of our technical root already illustrates thelong way development has yet to go. The mostobvious advantage of plants is their cellular growthwhich is facilitated by the ion-selective and expand-able cell membrane. Replicating such a compliantand still powerful device will remain a challenge forthe next decade. In consequence, the technical rootwill also not achieve lateral growth which is impor-tant for continuously counteracting any creeping ofthe substrate as a consequence of, e.g., externalloads.

In addition to that, the cellular growth of roottissue allows for almost friction free actuation. Ourtechnical concept works rather like an inverted tele-scope and hence displays reduced friction but notto the extent a plant root is able to achieve. The

10

simplicity of the plant sensors and control architec-ture is again unmatched and more research on thesesensors is necessary before they are completely un-derstood and await technical transfer. However,the potential of technical roots is obvious and ifbiomimetic research is continued we might one daymake use of these roots not only to anchor, but alsoto exploit existing resources for growth and exter-nal use.

Acknowledgment

This work has been supported by the Ariadnascheme of the European Space Agency (Study No.06/6301).

References

[1] Dario, P., Laschi, C., Mazzolai, B., Corradi,P., Mattoli, V., Mondini, A., Mancuso, S.,Mugnai, S., Masi, E., Azzarello, E., Hlavacka,A., and Pandolci, C., “Bio-inspiration fromPlants’ Roots,” Tech. Rep. 06/6301, EuropeanSpace Agency, Advanced Concepts Team,Available online at www.esa.int/act.

[2] Eapen, D., Barroso, M., Ponce, G., Cam-pos, M., and Cassab, G., “Hydrotropism: rootgrowth responses to water,” Trends Plant Sci.,Vol. 10, 2005, pp. 44–50.

[3] Esmon, C., Pedmale, U., and Liscum, E.,“Plant tropisms: providing the power of move-ment to a sessile organism,” Int. J. Dev. Biol.,Vol. 49, 2005, pp. 665–674.

[4] Baluska, F., Mancuso, S., Volkmann, D., andBarlow, P., “Root apices as plant commandcentres: the unique brain-like status of theroot apex transition zone,” Biologia, Vol. 59,2004, pp. 1–13.

[5] Thellier, M., Desbiez, M., Champagnat, P.,and Kergosien, Y., “Do memory processes oc-cur also in plants?” Physiol. Plant., Vol. 56,1982, pp. 281–284.

[6] Knight, H., Brandt, S., and Knight, M., “Ahistory of stress alters drought calcium sig-nalling pathways in Arabidopsis,” Plant J.,Vol. 16, 1998, pp. 681–687.

[7] Trewavas, A., “Aspects of plant intelligence,”Ann. Bot., Vol. 92, 2003, pp. 1–20.

[8] Misra, R., Dexter, A., and Alston, A., “Maxi-mum axial and radial growth pressures of plantroots,” Plant and Soil , Vol. 95, 1986, pp. 315–318.

[9] Bengough, A., Bransby, M., Hans, J.,McKenna, S., Roberts, T., and Valentine, T.,“Root responses to soil physical conditions;growth dynamics from field to cell,” J ExpBot , Vol. 57, 2006, pp. 437–447.

[10] NASA, “http://solarsystem.jpl.nasa.gov/planets/profile.cfm?Object=Mars,” 2008.

[11] Anttila, M., “Concept evaluation of Marsdrilling and sampling instrument,” Tech. rep.,Helsinki University of Technology, Laboratoryof Space Technology, March 2005.

[12] ESA, “Exomars rover/Pasteur System Re-quirements Document,” 2003.

[13] ESA, “Statement of Work for the Phase AStudy of the Pasteur Exobiology Payload andRover for the EXOMARS Mission,” 2003.

[14] NASA, “Odyssey Finds Water Ice in Abun-dance Under Mars’ Surface,” NASA Press Re-lease 02-99, May 28 2002.

[15] NASA, “http://www.nasa.gov/phoenix,”2008.

[16] Clinton, D., Holt, A., Jantz, E., Kaufman, T.,Martin, J., and Weber, R., “A Lunar/Martiananchor emplacement system. (Idaho Univ.),NASA-CR-197193,” Tech. rep., 1993.

[17] Bar-Cohen, Y., Chang, Z., Sherrit, S., Bade-scu, M., and Bao, X., “The Ultrasonic/SonicDriller/Corer (USDC) as a Subsurface Drill,Sampler and Lab-On-A-Drill for Planetary Ex-ploration Applications,” Proc. of the SPIESmart Structures Conference, San Diego, CA.,Vol. SPIE Vol. 5762-22, 2005.

[18] Clark, L., Whalley, W., and Barraclough, P.,“How do roots penetrate strong soil?” Plantand Soil , Vol. 255, 2003, pp. 93–104.

11

[19] Bengough, G., Croser, C., and Pritchard, J.,“A biophysical analysis of root growth undermechanical stress,” Plant and Soil , Vol. 189,1997, pp. 155–164.

[20] Pritchard, J., “The control of cell expansion inroots,” New Phytol , Vol. 127, 1994, pp. 3–26.

[21] Probine, M. and Preston, R., “Cell growth andthe structure and mechanical properties of thewall in internodal cells of Nitella opaca. II. Me-chanical properties of the walls,” J. Exp. Bot.,Vol. 13, 1962, pp. 111–127.

[22] Lockhart, J., “An analysis of irreversible plantcell elongation,” J Theor Biol., Vol. 8, 1965,pp. 264–275.

[23] Greacen, E. and Oh, J., “Physics of rootgrowth,” Nature New Biol., Vol. 235, 1972,pp. 24–25.

[24] Atkins, P., Physical Chemistry , Oxford Uni-versity Press, 1997.

[25] Su, Y. and Lin, L., “Water-Powered MicroDrug Delivery System,” Journal of Microelec-tromechanical Systems, , No. 13, 2004, pp. 75–82.

[26] Muday, G., Peer, W., and Murphy, A., “Vesic-ular cycling mechanism that control auxintransport polarity,” Trends Plant Sci., Vol. 8,2003, pp. 301–304.

[27] Friml, J., “Auxin transport shaping theplant,” Curr. Opin. Plant Biol., Vol. 6, 2003,pp. 7–12.

[28] Samaj, J., Baluska, F., and Hirt, H., “Fromsignal to cell polarity: mitogen-activated pro-tein kinases as sensors and effectors of cy-toskeleton dynamicity,” J. Exp. Bot., Vol. 55,2004, pp. 189–198.

[29] Liscum, E., “Phototropism: Mecha-nisms and outcomes,” The Arabidop-sis Book , 2002, American Society ofPlant Biologists, Rockville, Maryland,http://www.aspb.org/publications/arabidopsis.

[30] Darwin, C. and Darwin, F., The Power ofMovement in Plants, John Murray, London,1880.

[31] Blancaflor, E. and Masson, P., “Plant grav-itropism: unraveling the ups and downs ofa complex process,” Plant Physiol., Vol. 133,2003, pp. 1677–1690.

[32] Mugnai, S., Azzarello, E., Masi, E., Pandolfi,C., and S., M., “Nutation in plants,” Rhythmsin Plantss, Sprinter-Verlag, Berlin, 2007.

[33] Brown, A. H., “Circumnutations: From Dar-win to Space Flights,” Plant Physiol., Vol. 101,1993, pp. 345–348.

[34] Larson, K. C., “Circumnutation of an exotichoneysuckle vine and its native. Congener:influence on clonal mobility,” Am. J. Bot.,Vol. 87, 2000, pp. 533–538.

[35] Perbal, G. and Driss-Ecole, D., “Mechan-otransduction in gravisensing cells.” TrendsPlant Sci., Vol. 8, 2003, pp. 498–504.

[36] Babiasz, J., Cesarini, A., Gauby, S., Kovnat,S., and Stirling, R., “Autonomous AnchoringSystem for Exploratory Devices on Mars. Se-nior Design Project, Mechanical Engineeringand Mechanics Department, Drexel Univer-sity, USA.” .

12