Research Of Balance Control For Bionic Kangaroo Robot During Stance Phase

Hopping robot, with the advantage of its moving way, can pass through some discontinuous topography, such as ravines and rivers. The ability to adapt to the terrain of hopping robot has greatly improved compared to the wheeled robot and crawler robot, which can be better applied to emergency rescue, disaster relief and military reconnaissance in the harsh environment.In this paper, a new type of bionic kangaroo robot was designed, a study of the balance control of the standing bionic kangaroo robot was made. Detailed content of the work and results are as follows:1) The design of the bionic kangaroo robot systemReferred to the physical appearance of kangaroo, the structure and function of legs and tail, the mechanical structure of the bionic kangaroo robot were designed. The legs consists of two quadrilateral components, simulating the leg tendon function of kangaroo. The shape of the robot's tail modelled the shape of kangaroos' tail with a hollow structure, which reduces the tail moment of inertia effectively and the driving torque for driving the tail to rotate. The electronic control system based on STM32 was designed. This system realized the acquisition and processing of data, the driving control of the jumping mechanism as well as the display of the host computer system.2) Dynamic modeling for the bionic kangaroo robot during stance phaseFirstly according to the structure characteristics of the designed bionic kangaroo robot, we simplified mechanism model of the robot that standing on the ground. Since the robot's legs are coaxial synchronous, and its body is symmetrical with respect to the front, the robot mechanism can be simplified to single leg jumps of planar mechanism. By analyzing the structure of the robot, a multi-rigid body model of this robot standing on the ground is established. Secondly, a dynamics model of the multi-rigid body model of this robot was established by using Lagrange method. The system is a nonlinear system. Its state space model was derived after linearization.3)Balance control for the bionic kangaroo robot during stance phase based on LQR and HPSOA linear quadratic regulator for stance balance control is designed based on the dynamics model, in which the optimum weight matrix is obtained by hybrid particle swarm algorithm. Simulations are conducted on balance control of the robot during stance using the optimized LQR regulator. Results show that the optimized control method can improve the control performance of the bionic robot.4) Adaptive sliding mode control of the bionic kangaroo robot during stance phaseBecause of uncertain factors such as external disturbance, the actual control model and the dynamics model of the theoretical analysis may be different in the process of balance control for bionic kangaroo robot during stance phase. In order to solve this problem, a sliding mode controller based on adaptive exponential reaching law was designed. To improve traditional exponential reaching law, the adaptive principle was used in the design of reaching law. This design not only ensures the system moving point has faster convergence speed when away from the switching surface, but also weaken the chattering phenomenon. The simulation proves that the sliding mode controller based on adaptive exponential reaching law has a better effect to control.