Gait Synthesis and Modulation for Quadruped Robot Locomotion Using a Simple Feed-Forward Network

Abstract

Abstract. This paper describes a technique for statically stable gait synthesis for a quadruped robot using a simple Feed Forward Neural Networks (FFNN). A common approach for gait synthesis based on neural networks, is to use an implementation with Continuous Time Recurrent Neural Network (CTRNN) of arbitrary complex architecture as pattern generator for rhythmic limb motion. The preferred training method is implemented using genetic algorithms (GAs). However, to achieve the desired trajectory becomes an obstacle during the training process. This paper presents a much more simpler process converting a statically stable gait into actuator's space via inverse kinematics; the training of the network is done with those references. By doing so, the training problem becomes a spatio-temporal machine learning problem. It is described a solution for trajectory generation combining a simple oscillator model with a Multilayer Feedforward Neural Network (MFNN) to generate the desired trajectory.

Through the development of a biomechatronic knee prosthesis for transfemoral amputees: Mechanical design and manufacture, human gait characterization, intelligent control strategies and tests

“…In such sense, we asked the question, "What if a reference of knee angle is generated to be followed by the prosthesis?". The answer was found in central pattern generators (CPGs), which have been broadly used in the last decade to generate and coordinate motion of legs in walking robots, for executing all gait modes physically possible [16], [17]. Of course, there is a difference because the prosthesis is not a robotic device, designed to behave according to the programmer's orders but to the patient's will.…”

Cybernetic knee prosthesis: application of an adaptive central pattern generator

“…The Mechatronics Research Group (MERG) at the Simón Bolívar University, Caracas-Venezuela, has been working in the control of legged robot platforms; more specifically, in bio-inspired modeling of gait generation system in quadrupeds Cybernetic knee prosthesis and hexapods (Cappelletto, 2007). They have proposed a CPG architecture, based on separation of the gait generation system, into a temporal reference subsystem and a space transformation subsystem (Cappelletto et al, 2006). In this approach, there is a decoupling between the oscillating subsystem and the robot leg kinematics.…”