Literature

We know from personal experience that moving through a wood thick with undergrowth is a far more daunting task than walking across a grassy field. Imagine what it is like for arthropods!

"When animals, such as spiders and cockroaches, scurry through their natural environment, debris can decrease the probability of a foothold, provide diverse asperities for contact and flow like a fluid causing the animal to slip. Effective mobility over varied terrestrial substrates in the natural world that differ in geometry, compliance and tendency to flow must require a feedback component. Thus, control strategies of legged locomotion in nature cannot be determined without regard to an animal's environment."

What, then, are these control strategies? The first strategy to consider is distributed neural feedback. This requires active sensing by the animals followed by processing of those sensory signals. To do this, the organism must employ an "enormous number and rich variety of motion, contact, length and stress sensors." But it can only operate slowly. "Arthropods can negotiate obstacles, ditches, gaps and uneven surfaces, as well as travel up and along inclines by slow, deliberate stepping." This strategy does not work well when the animal wants to run.

"Highspeed locomotion can be advantageous on irregular terrain because kinetic energy allows the organism to bridge gaps in footholds that slow-moving systems find impassible. Rapid running, however, is a high bandwidth behavior. Delay in neural communication channels reduces synchronization gains, so that an animal's nervous system tends to operate in a decentralized, feedforward manner where coordination is achieved primarily through mechanical coupling and stability through preflexes."

This brings us to the second strategy: distributed mechanical feedback. This is not speeded up neural feedback, but an integration of active and passive neural responses with the kinaesthetic senses of the animal.

"We propose that the control of locomotion on challenging terrain can be simplified during rapid running using distributed mechanical feedback. We demonstrate that animals are effective at traversing challenging terrain at high speeds by distributing the mechanical feedback over limbs moving in appropriate trajectories with components that generate passive responses to leg-surface contact events. Distribution of the mechanical feedback creates effective coupling with environments, and results from the synergistic operation of leg trajectory, leg configuration and attachment mechanism. The control algorithms are in effect embedded in the form of animal itself: control results from the properties of their parts, their morphology and their passive interaction with the environment."

This perspective enables the organisms to understood in new ways:

"arthropods have an impressive array of attachment mechanisms on their feet that can increase a leg's probability of surface contact. Mechanisms include hooks or claws, suckers, glue and friction. Distal tarsal claws have been shown to increase performance on rough, inclined surfaces and during inverted locomotion. Mechanisms in some orientations can require feedback from the nervous system to engage, whereas others operate by passive mechanical feedback to respond to specific mechanical events."

This analysis has inspired research into robot control mechanisms. "The present study tests the hypothesis that distributed mechanical feedback simplifies the control of animal - and robot-surface interactions." The researchers selected substrates that they could define and use in simulation studies. The experimented with spiders, cockcroaches and crabs - multilegged runners. "We verified our results in a physical model, a rapid running, six-legged robot named RHex. We altered leg configuration and attachment mechanism, but not its electronic control strategy."

RHex has extraordinary rough-terrain mobility

This is from their conclusion:

"We found no evidence that these multilegged runners use neural feedback to follow a foreleg's secure foothold or to adjust the location of where their legs contact the substrate. By using the kinetic energy of rapid running to bridge gaps in footholds and distributing mechanical feedback over many legs and locations along the leg, animals can overcome the inherent delays of neural feedback as well the problem of noisy sensors, thereby simplifying control. Use of a physical model, a legged robot, supported our contention because the hexapod was able to traverse low probability foothold terrain better with a broader leg contact area and collapsible spines, but without a single change to its electronic controller. Our discovery of distributed mechanical feedback provided biological inspiration to a robot that can now traverse terrain previously impassable."

It is worth standing back and reflecting on what is happening here. The research team has devoted much time to understanding the control strategies of arthropods and has found that these organisms must be understood in a holistic way. The control system involves distributed neuronal and mechanical feedback mechanisms which are so sophisticated that our best efforts at emulation can only be described as crude. The robot RHex is undoubtedly a product of some highly intelligent research scientists and engineers, but why is it that people baulk at the thought that the multilegged runners they have been studying are also the product of intelligent design?

Distributed mechanical feedback in arthropods and robots simplifies control of rapid running on challenging terrain
J C Spagna, D I Goldman, P-C Lin, D E Koditschek and R J Full
Bioinspiration & Biomimimetics, 2(1) (March 2007) 9-18 | doi 10.1088/1748-3182/2/1/002

Abstract. Terrestrial arthropods negotiate demanding terrain more effectively than any search-and-rescue robot. Slow, precise stepping using distributed neural feedback is one strategy for dealing with challenging terrain. Alternatively, arthropods could simplify control on demanding surfaces by rapid running that uses kinetic energy to bridge gaps between footholds. We demonstrate that this is achieved using distributed mechanical feedback, resulting from passive contacts along legs positioned by pre-programmed trajectories favorable to their attachment mechanisms. We used wire-mesh experimental surfaces to determine how a decrease in foothold probability affects speed and stability. Spiders and insects attained high running speeds on simulated terrain with 90% of the surface contact area removed. Cockroaches maintained high speeds even with their tarsi ablated, by generating horizontally oriented leg trajectories. Spiders with more vertically directed leg placement used leg spines, which resulted in more effective distributed contact by interlocking with asperities during leg extension, but collapsing during flexion, preventing entanglement. Ghost crabs, which naturally lack leg spines, showed increased mobility on wire mesh after the addition of artificial, collapsible spines. A bioinspired robot, RHex, was redesigned to maximize effective distributed leg contact, by changing leg orientation and adding directional spines. These changes improved RHex's agility on challenging surfaces without adding sensors or changing the control system.

See also:

RHex Web site

RHex robot (video)

Feedback on this blog from an engineer working in robotics: "Seeing design in an animal as the result of intelligence tends to make me strive for deeper answers to the technical problems of bio/robomechanics and control, whereas seeing animals as the result of evolution tends to make thinking, in my opinion, much shallower."