Investigating the mechanism of whole-body motor control of lizard locomotion and its link to applications in robotics

It is said that the common ancestor of quadrupeds walked by flexing its trunk like a lizard. The research group of Professor Akio Ishiguro, Associate Professor Takeshi Kano, Dr. Shura Suzuki (PhD student at the time of the study) at the Research Institute of Electrical Communication (RIEC), Tohoku University, and Professor Auke J. Ijspert of the Biorobotics Laboratory, École Polytechnique Fédérale de Lausanne, Lausanne, Switzerland clarified the control mechanism by which the lizard-like gait produces whole-body movement. The researchers revealed a very simple control rule that "the limbs and trunk coordinate to assist each other's movements." In a world first, the group also succeeded in reproducing the change in the lizard-like walking pattern according to the moving speed. It is expected that these findings will lead to the deciphering of the mechanisms by which quadrupeds skillfully manipulate and coordinate the motion of their whole bodies and the development of new robots. The group published their findings online in Frontiers in Neurorobotics.

There are a wide variety of quadrupeds on Earth, but one point they all have in common is that they move by utilizing the movement of the whole body including their four legs, head, tail, and torso. If the mechanism that produces the whole-body movement common to quadrupeds can be clarified, it will deepen the biological understanding of the excellent motor ability of these animals. It may also be equally applied to engineering as a control measure for highly mobile robots able to move freely in the natural environment.

In lizard-like locomotion, the gait of limb and trunk movements change depending on conditions such as speed. This means that it is conceivable that there are mechanisms behind this behavior that flexibly allows the animal to alter whole-body movements according to their requirements. Another factor is that since the neural circuitry of the organisms that exhibit a lizard-like gait is simpler than that of mammals, such as dogs and cats, this gait might be achieved through a relatively simple motor control mechanism when compared to the behavior of other limbed animals. In addition, it has been clarified from footprint fossils, etc. that the ancestral quadruped (the common ancestor of extant limb animals) also used lizard-like locomotion. It is considered that this behavior contains an inherent essence of motor control that is common to all quadrupeds. These facts have suggested the importance of lizard-like locomotion in clarifying systemic motor control. However, it has not been clarified to date what kind of control mechanism is used to realize a lizard-like gait.

The research group created a simulation to clarify the mechanism by reproducing the movement of salamanders. They chose this method because is easy to measure information about nerve function and body movement in a simulation, and reproducible results can be obtained under various experimental conditions. In their experiments, the researchers used a control algorithm in a robot and confirmed its behavior. The robot consisted of a leg with two joints connected to the trunk with 10 joints.

The algorithm consisted of feedback rules which were broadly divided into three types of control mechanisms, responsible for 'limb coordination', 'trunk coordination', and 'limb-trunk coordination'. The functions of each mechanism are as follows: "coordination control of the limb," in which the leg continues to support the body while the leg is loaded; "coordination control of the trunk," in which each joint angle follows the joint angle of the front; and "coordination of the limb and trunk," in which the stride length becomes longer in proportion to each other part's movement.

The results of the simulation experiment confirmed that the group had developed a stable lizard-like gait. In addition, it was revealed that limb and trunk movements were altered depending on parameters such as locomotion speed and body structure. These changes in gait patterns have been confirmed to be largely consistent with actual animal behavior, supporting the validity of the proposed control algorithm.