Development And Research For Control System Of Humanoid Robot

Research on humanoid robot integrated plenty of high-new technology. The more robots like human, the more complicated control circuit is. How to optimize circuit's configuration and to reduce size and power dissipation of circuit in order to combine control circuit with mechanism more perfectly has already been one of important topics in research domain of humanoid robots. Humanoid robot is a multi-variable coupled nonlinear time-varying system, so conventional control algorithm can not satisfy the need of control. Many researchers apply themselves to intelligent control strategy.Firstly, the dimension and the DOF configuration of humanoid robot HBUT-1, the controlled object, were introduced. Mechanical parameters of robot joint were analyzed. The integral mechanical configuration of humanoid robot was introduced. Secondly, maxon DC motor was chosen to drive robot's joint. The dynamic mathematical model of maxon DC motor was made out. Thirdly, the integral configuration of humanoid robot's sensing system was introduced. Applications of digital vidicon, SONAR, 6-DOF force/torque sensor and opto-electronic encoder in robot's sensing system were introduced in detail. Fourthly, a type of humanoid robot control system with 3-layers frame composed of embedded single-board computers was introduced in detail. The control system distributes tasks to all embedded computers among 3 layers, so that the intelligent degree of robot is enhanced. Fifthly, the hardware configuration and the control strategy of a kind of joint controller for humanoid robot were introduced. The joint controller based on DSP TMS320F2811 features: full digital control, high integration level, small size, low power dissipation, multi-axis motion control, having CANBUS interface. A new control strategy which combined fuzzy PD control with adaptive control based on single neuron was used in the joint controller. Experiments showed that this control strategy could improve response speed, control precision and robustness. Sixthly, the gait optimization of humanoid robot was made. The humanoid robot HBUT-1 was simplified into a seven-link model with 12 rotational DOF. The vector describing robot's position and pose was established, then the vector's expected locus during a regular step was modelled by the 5th order polynomials. The mathematic descriptions of geometry restriction and stabilization were analyzed respectively, and then the optimal gait was worked out with genetic algorithm mixing binary number encoding and floating point number encoding. Experimental results show that the optimal gait maximizes dynamic stabilization.