The Walking Control of a Crab Biorobot in Amphibious Environment

Abstract

This article describes development of a crab biorobot that is capable of traversing diverse environments including both land and water. We have transformed a living rainbow crab into a walking biorobot by attaching wireless controller. An anatomical and physiological investigation revealed the rainbow crabs have sensory system on the carapace. Based on this finding, we implanted electrodes into the carapace. The walking direction of the robot is controlled through electrical stimulation provided by the controller. Depending on this site, the crab biorobot is induced to walk forward, leftward, and rightward in both terrestrial and underwater conditions. There is no significant difference in the mean walking direction between the two conditions. Smooth transition from land to water of the crab biorobot further demonstrates the adaptability to amphibious environment. This biorobot is compact, measuring 5 cm in carapace and weighing 50 g including the wireless controller. The crab biorobot in this scale has a potential for application narrow and unstructured in waterfront environments.

Introduction

Amphibious robots face locomotion challenges due to the physical differences of air and water. For example, conventional terrestrial robots employ wheels and tracks as their means of propulsion, relying on continuous surface contact and friction to generate thrust. This reliance on surface contact limits the mobility of wheeled and tracked robots in water, that is, in low gravity conditions.¹

As an alternative approach, more attention has been directed to the propulsion mechanism inspired by animal locomotion, such as walking,^2,3 undulation,^4,5 and flapping.^6,7 However, these robots tend to be bulky due to their complex mechanisms. There is a growing demand for small robots that can operate effectively in narrow and complex amphibious environments, including tasks such as pipeline inspection, reconnaissance, and disaster relief.

A biorobot, also known as an animal cyborg, is a hybrid composed of living animal and electronic components, employed to induce desired locomotion to the animal. This concept provides a viable solution for creating centimeter-scaled robots. Numerous studies have demonstrated that various animal species can serve as a scaffold of biorobots.⁸ Cockroaches^9–15 and beetles¹⁶ are often used to build up terrestrial walking biorobot, whereas beetles^17–19 and moths^20,21 can serve as platform of aerial biorobots.

Locusts^22,23 have been transformed into terrestrial robots by leveraging their unique mode of locomotion, jumping. In a recent study, jellyfish were used to create a swimming robot.²⁴ However, current biorobots are engineered to function in specific environments, such as on land, in water, or in the air. To date, there has not been a biorobot capable of traversing multiple environments.

Semiaquatic crabs exhibit remarkable locomotor ability both on land and in water, moving across solid ground, rocky area, sand, muddy field, and underwater in their natural habitat. When animals move from the land to water, its effective weight is reduced due to upward buoyancy force, whereas lift and drag are increased 800-fold. These differences lead to difficulty in legged locomotion in water, for example, falling and overturning.²⁵ To overcome the underwater locomotion problems, crabs employ a distinctive gait known as punting.^25–27 Hence, semiaquatic crabs can be considered as a suitable candidate for platform of amphibious biorobot due to their adaptability to terrestrial and aquatic environments.

Most of biorobots are equipped with electronic devices that manipulate the locomotion of the platform animal, allowing operators to guide the robots in their desired direction. This is achieved through the application of electrical stimulation to the muscles or nervous system.^{8,10–21,28,29} For example, cockroaches walk forward by response to electrical stimulation applied to their cerci, the sensory organs located at the end of their abdomen.⁹ To induce right or left turns in a cockroach, electrical stimulation is delivered to the left or right antenna, respectively.^10–13

Although these methods are effective to several insect species,^10–13,16 they are not applicable to crabs due to their distinctive body structure. Because crabs have small and folded antenna that fit to the body, antenna is not suitable as a site of electrode implantation. In addition, their abdomens fold beneath the thorax, making it difficult to access to the central nervous system on the ventral side.

In this article, we developed a amphibious biorobot utilizing the rainbow crab (Cardisoma armatum). First, we conducted anatomical and physiological investigation to pinpoint the site of electrical stimulation. Rainbow crabs were then equipped with wireless controller to manipulate their locomotion. Electrical stimulation to the crab resulted in leftward, forward, and rightward walking depending on the stimulation site. Finally, we demonstrated that the crab biorobot is capable of traversing an amphibious environment.

Materials and Methods

Animals

Male rainbow crabs (C. armatum) were used for all experiments. The animals were maintained in a plastic tank at 25°C and fed with dried shrimp ad libitum. Based on the guideline of the Nanyang Technological University Institutional Animal Care and Use Committee, our study did not require ethical approval as it involved crabs, which are invertebrates.

Anatomy

The crabs were collected from the tank, chilled on ice for 30 min, and fixed with 1% paraformaldehyde (PFA) solution at 4°C for 2 days. After fixation, the specimens were washed with crab saline (NaCl 374 mmol/L, KCl 7.5 mmol/L, CaCl₂ 15.4 mmol/L, MgCl₂ 9.9 mmol/L).³⁰ To expose the thoracic ganglion, the carapace was opened using the miniature electric drill and the internal organs were removed using tweezers and microscissors. Methylene blue was applied to enhance the contrast during the dissection.

Neural recording and mechanical stimulation

The crabs were collected from the tank and chilled on ice for 30 min before experiments. The chelipeds and walking legs were tied with rubber bands. The surface of the thoracic sterna was roughened using a miniature electric drill and a plastic holder was attached. The plastic holder was then fixed to a magnet base with screw so that the crab was harnessed dorsal side up (Fig. 1B). A small opening was made on the carapace to expose the carapace nerve near the surface.

FIG. 1.

Anatomy and physiology of the sensory system on the carapace. (A) Top panel: dorsal view of a dissected rainbow crab. The carapace was opened from the dorsal side, and the internal tissues were removed to expose the thoracic ganglion. The blue and green dotted squares indicate the area magnified in middle and bottom panels, respectively. Middle panel: a pair of nerve cord emerges from the thoracic ganglion (arrowheads). The nervous system was labeled with methylene blue for visualization. Bottom panel: The nerve cord shown in the middle panel is divided near the carapace surface and projects into the epidermal layer. Neural activities were recorded from the nerve cord proximally to the branching point (asterisk). (B) Neural recording setup. Neural activities were recorded from the nerve cord indicated in A. The custom-made strain gauge was attached to the center of the carapace, whereas the accelerometer was attached to the actuator tip. (C) Example of neural recording. Top trace: neural activity. Bottom trace: output signal from the strain gauge (filtered with Butterworth bandpass filter). The spike discharges were observed in the middle part of the cycle of vibration. (D) Raster plot and peristimulus time histogram of neural response. Neural response showed clear hysteresis, that is, spike discharged was more prominent in ramp-up phase than in ramp-down phase. (E) Correlation between firing rate and vibration amplitude. Blue points: firing rate during the ramp-up phase. Orange points: firing rate during the ramp-down phase. Blue line: linear regression-fit of firing rate to vibration amplitude during the ramp-up phase. Orange line: linear regression-fit of firing rate to vibration amplitude during ramp-down phase.

The glass suction electrodes were used to record the neural activities from the nerve cord. The tip of the glass pipettes was cut and blunt by flaming and, the shaft was filled with crab saline. The carapace nerve was cut using microscissors and the distal end of the nerve was sucked into the glass electrode. A reference electrode was put near the suction electrode. Ag-AgCl wire was used for both recording and reference electrodes. The neural signal was amplified (Model 3000; AM-Systems) and digitized through an Analog-to-Digital converter (PowerLab 26T; AD Instruments), then stored in the computer using commercial software (LabChart; AD Instruments).

To examine whether the carapace nerve contains sensory neurons, mechanical stimulus was delivered to the crab's body. A plastic cone was attached on the electromagnetic actuator and positioned above the crab that was secured on the magnet base. The tip of the plastic cone was fixed to the center of the carapace (Fig. 1B). The actuator was driven by a function generator (33500B; KEYSIGHT). The function generator was triggered by a custom-made software through a microcomputer, which then sent an amplitude-modulating sine wave to the actuator. The carrier frequency was set at 200 Hz and the amplitude was modulated at 0.1 Hz. The magnitude of the vibration was monitored using the custom-made strain gauge.

Analysis of neural activity

Neural signals were filtered with a second order Butterworth bandpass filter (250–3000 Hz), and spike activities were detected by amplitude thresholding. Gaussian smoothing was used to estimate the firing rate. The spike activities were binned with 1 ms window and convolved with Gaussian kernel. Then, firing rate was calculated by averaging the smoothed spike activity in every 100 ms time window.

Sensitivity of the custom-made strain gauge was calibrated using the accelerometer at 50 Hz. The amplitude of vibration applied to the crab was calculated based on the output signal of the strain gauge.

Wireless stimulation device and stimulus parameter

The crab biorobot was controlled by a customized wireless circuit board (Fig. 2B). Texas Instruments (TI) MSP432P4011 microcontroller (ARM 32-bit Cortex M4F, 48 MHz, 2 MB of Flash, 256 kB of SRAM) was used as the system main controller unit for its excellent power efficiency. The stimulation signal was generated by the chip AD5504 (Quad-channel, 12-bit, 7.3 mV resolution; Analog Devices) with voltage supplied up to 12 V using TI TPS61046 Boost Converters. The crab biorobot system was controlled wirelessly through Bluetooth 5.1 (2.4 GHz) using CC1352 microcontroller unit (ARM 32-bit Cortex M4F, 48 MHz, 352 kB of Flash, 8 kB of SRAM).

FIG. 2.

Overview of crab biorobot system. (A) As an amphibious animal, the crab biorobot can traverse multiple terrains, including grassland, rocky areas, and waterfront in the field. (B) The crab biorobot consists of a living rainbow crab (Cardisoma armatum) and a wireless stimulator (i.e., a backpack controller). Four electrodes are implanted into the crab's body to execute its locomotion control. (C) The crab biorobot performs three fundamental motions through electrical stimulation. The stimulation of two left/right electrodes induces a rightward/leftward walk, whereas the electrical activation of the two bottom electrodes elicits forward motion. (D) The control diagram of the amphibious biorobot system. The backpack controller allows operators to wirelessly maneuver the crab biorobot through a GUI. After being transferred from the GUI to the backpack, the maneuvering commands (i.e., leftward, rightward, or forward walk) are converted to electrical stimuli activating the crab's sensory systems. The crab employs this artificially provided information together with those it collected from the surrounding environment to alter its locomotion. GUI, graphical user interface.

The lithium–polymer battery was mounted on the backpack and whole body was coated with silicone sealant for waterproofing. Two small plastic spheres were attached on the assembled backpack/battery as markers to track the position of the crab (Supplementary Fig. S1C). The weight of the backpack was 6.0 g including the battery, markers, and sealant. The walking velocities were compared before and after mounting the backpack on the crabs and there was no significant difference (N = 4 crabs, n = 20 trials, for all experiment conditions; repeated measurement two-factor ANOVA, p > 0.05).

To stimulate the crabs, we used bipolar square wave with an amplitude of 3 V (zero-to-peak value), a pulse width of 12 ms (i.e., 41.7 Hz), and a duration of 2 s. The duty cycle was 50%.

Electrode implantation and attaching the backpack

Electrodes were made from Teflon-coated silver wire with 100 μm in diameter. Insulation was removed using a torch on both ends and one end of each wire was soldered to platinum wire (50 μm in diameter). The platinum wire was cut to 1 mm for implantation. The other end of silver wire was soldered to four-channel PCB header to interface with the backpack.

The crabs were immobilized by chilling on ice for 30 min. The surface of the central region of the carapace was roughened using an electric drill and a plastic plate (10 mm in length, 10 mm in width, and 1.5 mm in thickness) was fixed using superglue. The backpack was mounted to the plastic plate using double-sided tape right before the experiment.

We implanted four electrodes into the crab's body. The implantation site was determined based on the geometry of the crab carapace (Supplementary Fig. S1A). After the implantation procedure, the crabs were individually housed in a small plastic container and allowed to rest at least 1 day before the experiment.

Once the backpack was mounted on the crab, we applied beeswax to the battery connector and pins to seal any gap for waterproofing.

Behavioral experiment

Six rainbow crabs were used for stimulation experiment (46.7 ± 2.9 mm in carapace width and 52.6 ± 12.0 g in body weight). The crabs used in the behavioral experiment were with intact legs and chelipeds. All the animals had the major cheliped on the right side.

After mounting the backpack onto the crab, we released the crab in the experimental arena and recorded their behavior using a video camera (HDR-CX405; SONY, Japan) from 1330 mm above the arena (Supplementary Fig. S1B). The experimental arena was a large plastic container with 700 mm in length, 480 mm in width, and 200 mm in height. The floor was covered by a black neoprene rubber sheet. The arena was surrounded with a white paper to avoid any undesirable visual effect from the environment. The entire arena was illuminated with LED attached on the article. When we controlled the crab underwater, the arena was filled with fresh water to 5.5 cm depth so that the crab's entire body was submerged.

In each experiment session, the crab underwent stimulation up to 20 times through the same stimulation site. The amplitude and duration of electrical stimulation were set to 3 V_0-p and 2000 ms, respectively. Interval between the stimulation was at least 20 s. The stimulation site was randomly chosen before the session. The crab's locomotion was recorded for 6 s commencing 2 s before the onset of stimulation. The stimulation was manually triggered by the experimenter when the crab was located at the center of the arena. One of LEDs on the backpack synchronized with electrical stimulation applied to the crab.

Video analysis

We selected 15 trials from each session for analysis, ensuring that none of the crabs crashed against the wall. The locomotion of the crab was analyzed by tracking the position of two markers using MATLAB (Mathworks) and open-source software (DLTdv8).³¹ First, the markers were digitized using automatic tracking function in DLTdv8 and then manually corrected if needed. The anterior marker represented the position of the crab and the line that goes through two markers represented body orientation. The position of the top left corner of the backpack was also tracked to extract the onset of the stimulation by means of the brightness of the LED on the backpack (Supplementary Fig. S1C). The spatial resolution of the video frame was 0.5 mm per pixel and the frame rate was 25 frames/s.

In this article, we define the longitudinal axis by the imaginary line through the center of the carapace from posterior to anterior, whereas the lateral axis is defined by the imaginary line perpendicular to the longitudinal axis. The longitudinal and lateral axes are positive to above and left, respectively. Translational (longitudinal and lateral) walking velocities are calculated from the displacement and turning angle between consecutive video frames (Supplementary Fig. S1D, E). When the crab moves to coordinates (x, y) with respect to the original system, the crab's displacement along longitudinal axis (x′) and lateral axis (y′) in the coordinate system is: $x' = x c o s α + y s i n α,$ $y' = - x s i n α + y c o s α,$

where α is the turning angle. Longitudinal and lateral velocities are calculated by multiplying x′ and y′ by frame rate. The grand mean, which is the average of the means of translational velocities over 2-s window, is used for statistical analysis. The walking direction of the crab is calculated as the central angle of the coordinates with reference to the initial position (Supplementary Fig. S1E). We measured the walking direction in the 2-s window both before and during stimulation.

Statistical analysis

We conducted an omnibus test³² to examine whether spontaneous walking of the crabs was directed to a random direction, Kuiper test³² to examine whether the distribution of walking direction changes before and during electrical stimulation, and Watson–Williams test³² to examine whether mean walking direction is different among stimulation site and between on land and in water. For multiple comparison, p-value was adjusted using Bonferroni correction to control the occurrence of Type I error.

A one-way ANOVA was used to compare the walking speed or translational velocities before and during electrical stimulation. A one-way ANOVA was used to determine whether there are any significant differences between crabs in each experiment condition. Because there was no significant difference among crabs, data were pooled for each analysis.

All statistical analyses were performed on MATLAB software.

Results

Design and demonstration of crab biorobot

The overview of amphibious crab biorobot is shown in Figure 2. The crab biorobot consists of a rainbow crab and a wireless controller (backpack) (Fig. 2B). The backpack is equipped with four digital-to-analog output channels to deliver electrical stimulation to the crab. Four electrodes are implanted onto the crab's carapace and are connected to the output channels on the backpack. The electrical stimulation through electrodes can induce walking in the direction opposite to the stimulation site (Fig. 2C).

When the electrical stimulation is applied to the crab through a pair of the electrodes on the rear side, the animal elicits forward walking. Electrical stimulation through the left or right electrode pair can induce rightward or leftward walking, respectively. Figure 2D illustrates a schematic representation of the crab biorobot system. The crab's locomotion is controlled by operator using a custom-made software.

The crab biorobot is designed to function in an amphibious environment that requires the transition between land and water. As presented in Figure 3, we conducted a corresponding demonstration. The crab biorobot was released to the water-filled arena to perform a point-to-point navigation. The crab biorobot started from the platform above water surface, descended an incline on the right, walked in water through the gate at the center of the arena, and successfully reached the target at the left end (Fig. 3B and Supplementary Video S1). In the following sections, we describe the details of nervous system on the carapace and of the locomotion control through electrical stimulation.

FIG. 3.

Demonstration in amphibious environment. (A) An illustration of crab biorobot locomotion. The crab biorobot can walk on land, make transition to water, and approach the target passing through the gate submerged in water. (B) Superimposed image of the crab biorobot demonstration. The arena is 500 × 500 mm in size and filled with water to a height of 25 mm. The operator manually triggered electrical stimulation through GUI to navigate the crab biorobot along the predefined route. The amplitude and duration of electrical stimulation were set to 3 V and 1000 ms to improve maneuverability during point-to-point navigation.

Anatomy and physiology of the sensory system on the carapace

The carapace, which is the largest part of the crab's body and faces upward, offers easy access for experimenters. Inspired by the stimulation protocol used in cockroaches,³³ we considered that variations in the stimulation site on the carapace could have a significant impact on crab locomotion. However, there has been limited knowledge regarding the nervous system on the carapace.³⁴ Figure 1 shows anatomy and physiology of the neurons projected to the carapace. A pair of nerve cords emerges from the dorsal side of the thoracic ganglion and extends to the epidermal layer beneath the carapace (Fig. 1A).

Neural recording from the nerve cord revealed that crabs are sensitive to the mechanical stimulation on the carapace (Materials and Methods section, Fig. 1B–E). Spike activities were augmented when the vibration was applied to the carapace, and the firing rate showed a positive correlation with the vibration amplitude (ramp-up: r = 0.686, p < 0.001; ramp-down: r = 0.571, p < 0.001). Based on these findings, we chose four specific sites on the carapace for electrode implantation so that electrical stimulation activates the sensory system beneath and induces locomotor response to the crab.

Locomotion control of crab biorobot on land and in water

The rainbow crabs primarily walk sideways (Fig. 4 and Supplementary Figs. S1F and S2) on land as many of crab species do.^26,35 When the electrical stimulation was applied to the crab using the electrode pair on the rear side, the crab biorobot exhibited forward walking (middle column, Fig. 4). The electrical stimulation on the right side induced leftward walking and vice versa (left and right columns, Fig. 4). The direction of spontaneous and induced walking was significantly different for all stimulation sites (Kuiper test, Table 1).

FIG. 4.

Walking direction of spontaneous (upper panels) and induced (lower panels) walking on land. Polar plots present the distribution of direction toward which the crab biorobot walks in 2 s window before and during stimulation. Left column: crab biorobots were stimulated through right electrodes (N = 5 crabs, n = 75 trials). Middle column: crab biorobots were stimulated through rear electrodes (N = 5 crabs, n = 75 trials). Right column: crab biorobots were stimulated through left electrodes (N = 3 crabs, n = 45 trials).

Table 1.

Walking Direction of the Crab Biorobot in Terrestrial and Aquatic Conditions

	Terrestrial condition			Underwater condition
	Leftward walk	Forward walk	Rightward walk	Leftward walk	Forward walk	Rightward walk
Before stimulation	16.62 ± 0.86	33.23 ± 0.93	1.72 ± 1.71	12.61 ± 0.79	2.86 ± 1.11	16.62 ± 1.26
During stimulation	32.09 ± 0.46	7.45 ± 0.46	−36.10 ± 0.85	42.40 ± 0.42	1.15 ± 0.67	−40.11 ± 0.74
p	<0.01 (n = 75)	<0.01 (n = 75)	<0.01. (n = 45)	<0.01 (n = 90)	<0.01 (n = 60)	<0.01 (n = 60)

Values are mean ± standard error (degrees).

Furthermore, we confirmed that the direction of induced walking was significantly different across all stimulation sites (Watson–Williams test; left stimulation—rear stimulation: p < 0.01, left stimulation—right stimulation: p < 0.01, forward stimulation—right stimulation: p < 0.01).

The crabs walk sideways in water when they are not stimulated (Fig. 5 and Supplementary Fig. S2). In aquatic condition, direction of stimulus-induced walking was significantly different from spontaneous walking for all stimulation sites (Kuiper test, Table 1). The mean direction of induced walking was significantly different among all stimulation sites (Watson–Williams test; left stimulation—rear stimulation: p < 0.01, left stimulation—right stimulation: p < 0.01, rear stimulation—right stimulation: p < 0.01).

FIG. 5.

Walking direction of spontaneous (upper panels) and induced (lower panels) walking in water. Left column: crab biorobots were stimulated through right electrodes (N = 6 crabs, n = 90 trials). Middle column: crab biorobots were stimulated through rear electrodes (N = 4 crabs, n = 60 trials). Right column: crab biorobots were stimulated through left electrodes (N = 4 crabs, n = 60 trials).

There was no significant difference in direction of stimulus-induced walking between terrestrial and underwater conditions (Watson–Williams test; left stimulation: p > 0.05, rear stimulation: p > 0.05, right stimulation: p > 0.05). These results indicate that crab's response to electrical stimulation is consistent on land and in water.

Change in body orientation during stimulation

The mean walking direction during stimulation varied from −40.1° to 42.4° (Table 1). However, the crabs did not exhibit such substantial turning. The mean turning angle during stimulation ranged from −4.58° to 4.58° (Supplementary Fig. S3), and tended to show a positive correlation with the walking direction (Supplementary Fig. S4). These results suggest that crabs primarily alter their walking direction through adjustments in the ratio of translational movement, rather than turning.

Change in walking velocity during stimulation

Table 2 gives the translational walking velocities of the crab biorobots in both terrestrial and underwater conditions (Supplementary Figs. S5 and S6). As expected, the longitudinal and lateral velocities altered between spontaneous and stimulus-induced walking. The longitudinal velocity increased during stimulation in all conditions, whereas change in the lateral velocity was contingent on the stimulation sites. The mean walking speed significantly increased when the crabs were stimulated through rear and left electrodes in water (Supplementary Table S1).

Table 2.

Change in Longitudinal and Lateral Velocities Before and During Stimulation

	Longitudinal velocity (mm/s)
	Terrestrial condition			Underwater condition
	Leftward walk	Forward walk	Rightward walk	Leftward walk	Forward walk	Rightward walk
Before stimulation	3.1 ± 0.18	2.4 ± 0.13	2.9 ± 0.19	1.5 ± 0.09	5.8 ± 0.15	3.6 ± 0.15
During stimulation	10.0 ± 0.18	16.9 ± 0.18	11.3 ± 0.16	9.0 ± 0.16	18.8 ± 0.19	12.7 ± 0.18
p	<0.01 (n = 75)	<0.01 (n = 75)	<0.01. (n = 45)	<0.01 (n = 90)	<0.01 (n = 60)	<0.01 (n = 60)

	Lateral velocity (mm/s)
	Terrestrial condition			Underwater condition
	Leftward walk	Forward walk	Rightward walk	Leftward walk	Forward walk	Rightward walk
Before stimulation	−1.8 ± 0.28	−5.7 ± 0.40	0.7 ± 0.39	3.5 ± 028	1.0 ± 0.39	2.9 ± 0.34
During stimulation	10.5 ± 0.28	−4.3 ± 0.21	−9.2 ± 0.28	12.9 ± 0.28	−1.3 ± 0.27	−12.0 ± 0.19
p	<0.01 (n = 75)	n.s. (n = 75)	<0.01 (n = 45)	<0.01 (n = 90)	n.s. (n = 60)	<0.01 (n = 60)

Values are mean ± standard error.

n.s., nonsignificant.

Discussion and Conclusion

In this study, we have developed an amphibious biorobot utilizing a rainbow crab. This crab biorobot is equipped with a wireless controller and can execute three essential movements (forward, leftward, and rightward walking) through electrical stimulation in both terrestrial and underwater conditions. As a proof of concept, we conducted a demonstration of point-to-point navigation using the crab biorobot in an amphibious environment. The crab biorobot holds potential for various applications in narrow and complex amphibious environments, such as pipeline inspection, reconnaissance, and disaster relief.

Animals breathe to obtain oxygen for metabolization and to expel carbon dioxide. From the viewpoint of gas exchange, crabs are highly suitable as a platform of amphibious biorobot. Some invertebrate species, such as diving beetle³⁶ and aquatic spider,³⁷ are capable of transition between air and water. These animals use air store and gas gills, requiring them to return to the water surface to replenish bubbles. In contrast, crabs perform gas exchange using gills when in water, allowing them to remain submerged for long time without floating to the water surface.

However, the degree of terrestrial adaptation varies among crab species.³⁸ When aquatic blue crab (Callinectes sapidus) is exposed to the air, ventilation volume and oxygen extraction decrease, resulting in reduction of oxygen consumption to one-third of its aquatic value. In contrast, oxygen consumption of the terrestrial crab (Gecarcinus lateralis) diminishes to one-seventh of its aerial value when submerged in water. Only amphibious crabs maintain consistent oxygen consumption in both media. Therefore, amphibious rainbow crab is an ideal candidate for creating an amphibious biorobot.

As shown in Figure 1, we discovered that the neurons beneath the carapace are sensitive to mechanical stimulation. Electrical stimulation applied to the carapace can activate these neurons, leading to a locomotor response that causes the animal to move away from the source of stimulation. This suggests that electrical stimulation acts as a noxious stimulus for the crab. Although further investigations are necessary to uncover the underlying mechanism, it is reasonable to speculate that transient receptor protein (TRP) neurons play a pivotal role in the locomotor reaction, as TRP neurons are known to be involved in detecting a variety of stimuli, including chemicals, temperature changes, and mechanical stimulation.^39–41

Bluetooth operates as a low-power system with a range of ∼30 m in the air, but only a few inches in water. The backpack used in this study can communicate with the central station through Bluetooth up to 5 cm depth in water. When the crab biorobot is released to the experiment arena filled with water to a depth of 5.5 cm, the backpack was positioned at a depth of 2–3 cm in the water. Under this condition, communication between the backpack and the central station maintained stable throughout the experiment. For practical applications, it is essential to transmit signals over longer distances in water, which can be achieved through wireless technique such as acoustic communication.

Footnotes

Acknowledgments

The authors thank Dr. Long Duc Le, Dr. Phuoc Thanh Tran-Ngoc, and Chong Bing Sheng at School of MAE, NTU, for their helpful comments and advice. The authors also appreciate the support from Ms. Kerh Geok Hong Wendy for the coordination for the experiment.

Author Disclosure Statement

No competing financial interests exist.

Funding Information

This study was supported by the Singapore Ministry of Education (no. RG140/20) and Japan Science and Technology Agency under its JST-Mirai Program (no. JPMJMI21I1).

Supplementary Material

References

Inoue

, Shiosawa

, Takagi

. Dynamic analysis of motion of crawler-type remotely operated vehicles. IEEE J Ocean Eng, 2013; 38:375–382.

Greiner

, Shectman

, Won

, et al. Autonomous legged underwater vehicles for near land warfare. In: Proceedings of Symposium on Autonomous Underwater Vehicle Technology, 2 to 6 June 1996, Monterey, CA, USA. New York: IEEE; 2002; pp. 41–48.

Harkins

, Ward

, Vaidyanathan

, et al. Design of an autonomous amphibious robot for surf zone operations: Part II—Hardware, control implementation and simulation. In: Proceedings of the 2005 IEEE/ASME International Conference on Advanced Intelligent Mechatronics, 24 to 28 July 2005, Monterey, CA, USA. New York: IEEE; 2005; pp. 1466–1470.

Ijspeert

, Crespi

, Ryczko

, et al. From swimming to walking with a salamander robot driven by a spinal cord model. Science, 2007; 315:1416–1420.

Crespi

, Ijspeert

. Online optimization of swimming and crawling in an amphibious snake robot. IEEE Trans Robot, 2008; 24:75–87.

Baines

, Freeman

, Fish

, et al. Variable stiffness morphing limb for amphibious legged robots inspired by chelonian environmental adaptations. Bioinspir Biomim, 2020; 15:025002.

Baines

, Patiballa

, Booth

, et al. Multi-environment robotic transitions through adaptive morphogenesis. Nature, 2022; 610:283–289.

Webster-Wood

, Guix

, Nicole

, et al. Biohybrid robots: Recent progress, challenges, and perspectives. Bioinspir Biomim, 2023; 18:015001.

Tran-Ngoc

, Le

, Chong

, et al. Intelligent insect–Computer hybrid robot: Installing innate obstacle negotiation and onboard human detection onto cyborg insect. Adv Intell Syst, 2023; 5:2200319.

10.

Dirafzoon

, Bozkurt

, Lobaton

. Framework for mapping with biobotic insect networks: From local to global maps Robot. Auton Syst, 2017; 88:79–96.

11.

Latif

, Bozkurt

. Roach biobots: Toward reliability and optimization of control. IEEE Pulse, 2017; 8:27–30.

12.

Latif

, Whitmire

, Novak

, et al. Sound localization sensors for search and rescue biobots. IEEE Sens J, 2016; 16:3444–3453.

13.

Moore

, Crary

, Koditschek

, et al. Arbor A. Directed locomotion in cockroaches: “biobots.”. Acta Entomol Sloven, 1998; 6(2):71–78.

14.

Ariyanto

, Refat

CMM

, Hirao

, et al. Movement optimization for a cyborg cockroach in a bounded space incorporating machine learning. Cyborg Bionic Syst, 2023; 4:0012.

15.

Cole

, Bozkurt

, Lobaton

. Localization of biobotic insects using low-cost inertial measurement units. Sensors, 2020; 20(16):4486.

16.

Vo Doan

, Tan

MYW

, Bui

, et al. An ultralightweight and living legged robot. Soft Robot, 2018; 5:17–23.

17.

, Wu

, Sato

Feedback control-based navigation of a flying insect-machine hybrid robot. Soft Robot, 2018; 5; 365–374.

18.

, Sato

, Li

. Feedback altitude control of a flying insect–computer hybrid robot. IEEE Trans Robot, 2021; 37:2041–2051.

19.

Vo Doan

, Sato

. Insect-machine hybrid system: Remote radio control of a freely flying beetle (Mercynorrhina torquata). J Vis Exp, 2016; 115:e54260.

20.

Bozkurt

, Gilmour Jr

, Lal

. Balloon-assisted flight of radio-controlled insect biobots. IEEE Trans Biomed Eng, 2009; 56:2304–2307.

21.

Tsang

, Stone

, Aldworth

, et al. Flexible split-ring electrode for insect flight biasing using multisite neural stimulation. IEEE Trans Biomed Eng, 2010; 57:1757–1764.

22.

, Liu

, et al. Launching of a cyborg locust via co-contraction control of hindleg muscles. IEEE Trans Robot, 2022; 38:2208–2219.

23.

, Li

. The steering jump control of a locust bio-robot via asynchronous hindleg kickings. Adv Intell Syst, 2022; 4:2200082.

24.

, Townsend

, Costello

, et al. Field testing of biohybrid robotic jellyfish to demonstrate enhanced swimming speeds. Biomimetics, 2020; 5:64.

25.

Martinez

, Full

, Koehl

MAR

. Underwater punting by an intertidal crab: A novel gait revealed by the kinematics of pedestrian locomotion in air versus water. J Exp Biol, 1998; 201:2609–2623.

26.

Martinez

MM.

Running in the surf: Hydrodynamics of the shore crab Grapsus tenuicrustatus. J Exp Biol, 2001; 204:3097–3112.

27.

Chellapurath

, Stefanni

, Fiorito

, et al. Locomotory behaviour of the intertidal marble crab (Pachygrapsus marmoratus) supports the underwater spring-loaded inverted pendulum as a fundamental model for punting in animals. Bioinspir Biomim, 2020; 15:055004.

28.

Cao

, Zhang

, Choo

, et al. Insect-machine hybrid robot: Insect walking control by sequential electrical stimulation of leg muscles. In: Proceedings of the 2015 IEEE International Conference on Robotics and Automation, 26 to 30 May 2015, Seattle, WA, USA. New York: IEEE; 2015; pp. 4576–4582.

29.

Choo

, Li

, Cao

, et al. Electrical stimulation of coleopteran muscle for initiating flight. PLoS One, 2016; 11:e0151808.

30.

Taylor

, Greenaway

. Osmoregulation in the terrestrial Christmas Island red crab Gecarcoidea natalis (Brachyura: Gecarcinidae): Modulation of branchial chloride uptake from the urine. J Exp Biol, 2002; 205:3251–3260.

31.

Hedrick

Software techniques for two- and three-dimensional kinematic measurements of biological and biomimetic systems. Bioinspir Biomim, 2008; 3:034001.

32.

Zar

JH.

Biostatistical Analysis. 4^th ed. Prentice Hill: Upper Saddle River; 1999.

33.

Lin

, Zhang

, Kai

, et al. Resilient conductive membrane synthesized by in-situ polymerisation for wearable non-invasive electronics on moving appendages of cyborg insect. npj Flex Electron, 2023; 7:42.

34.

Kawamura

, Bagarinao

, Loke

, et al. Touch-sensitive bristles on the carapace of the mud crab Scylla paramamosain may be receptors for courtship signal. Fish Sci, 2021; 87:65–70.

35.

Blake

. Crab carapace hydrodynamics. J Zool, 1985; 207:407–423.

36.

Jones

, Seymour

. Gas exchange and dive behaviour in the diving beetle Platynectes decempunctatus (Coleoptera: Dytiscidae). J Insect Physiol, 2021; 133:104286.

37.

Seymour RS, Hetz SK. The diving bell and the spider: The physical gill of Argyroneta aquatica. J Exp Biol 2011; 214:2175–2181.

38.

O'Mahony

, Full

, Respiration of crabs in air and water. Comp Biochem Physiol, 1984; 79:275–282.

39.

Kozma

, Ngo-Vu

, Wong

, et al. Comparison of transcriptomes from two chemosensory organs in four decapod crustaceans reveals hundreds of candidate chemoreceptor proteins. PLoS One, 2020; 15:e0230266.

40.

Fernández-Carvajal

, Fernández-Ballester

, Devesa

, et al. New strategies to develop novel pain therapies: Addressing thermoreceptors from different points of view. Pharmaceuticals, 2012; 5:16–48.

41.

Himmel

, Letcher

, Sakurai

, et al. Drosophila menthol sensitivity and the Precambrian origins of transient receptor potential-dependent chemosensation. Phil Trans R Soc B, 2019; 374:20190369.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

3.25 MB

0.19 MB

0.06 MB

0.08 MB

0.05 MB

0.14 MB

0.13 MB

0.05 MB

0.00 MB