变幅轮腿机器人智能越障步态规划与平稳性分析

Stability analysis and gait planning for luffing wheel-legged robot during intelligent obstacle-surmounting process

为了实现机器人林区伐根的智能越障,保证机器人搭载扫描设备时的越障平稳性,该文设计了一种主动摆臂六轮腿式机器人结构,它具备2个独立的摆臂轮腿运动单元以及4个复合的摆臂轮腿运动单元。该文运用拉格朗日方程建立了机身智能越障过程中的摆臂轮腿动力学模型,采用最小二乘法拟合推杆的速度函数,通过ADAMS动态仿真以及样机试验,得到该变幅轮腿机器人在智能越障10 cm高度的过程中,机身的最大侧倾角与纵倾角较被动碰撞越障的右倾4.5°和前倾2.5°减小到左倾0.75°和前倾0.4°,验证了智能越障理论建模的正确性以及该系统的可靠性。该研究为该机器人在人工林区扫描作业的平稳越障提供了理论基础。

Luffing wheel-leg robot is applied to the operation on uneven surface, because of the characteristics such as high mobility, obstacle-surmounting capability and strong stability. In order to guarantee the smoothness of obstacle surmounting, the paper introduces a novel robot with 6 wheel-legs. This robot is a mobile equipment designed to surmount obstacles actively on forest road. The robot is a combination of 2 frameworks, 2 rear wheel-legs and 2 front wheel-legs. Wheel-legs are attached to the frameworks and distributed on both sides symmetrically. Linear actuators connected between wheel-legs and frameworks can lift the wheel-leg up and down through the obstacle-surmounting process. Front wheel-leg consists of a front straight wheel-leg and a front inverse V-shaped wheel-leg which means the front wheel-leg has 2 degrees of freedom. In order to achieve the intelligent obstacle-surmounting skill, 2 linear actuators are applied to the front wheel-leg on each side of the framework. Through analyzing the simplified model, the intelligent obstacle-surmounting process can be divided into 3 stages： obstacle surmounting of the first wheel in front wheel-leg, obstacle surmounting of the second wheel in front wheel-leg and obstacle surmounting of the rear wheel. The former 2 stages are controlled by the composite motion of the front wheel-legs, and the third one is controlled by the rear wheel-leg motion only, which means intelligent obstacle surmounting of the front wheel-leg is the key factor during the process. The kinematic model is established based on the movement relationship of the system. Through the calculation of the kinematic model, the maximum height of the intelligent obstacle-surmounting process is obtained, which is related to the length of the inverse V-shaped wheel-legs, the angle between the 2 legs in the inverse V-shaped wheel-legs and the swing angle range of the inverse V-shaped wheel-legs. According to the motion differential equation, the Jacobian matrix between the velocities on the wheel center of inverse V-shaped wheel-legs and the linear actuators is achieved. Based on the Lagrange equation, the dynamic model of the wheel-legs during the obstacle-surmounting process is established. According to the dynamic model, the force of the linear actuators can be calculated at the beginning of the progress. In ADAMS, the obstacle with 100 mm height and 180 mm diameter is established for the intelligent obstacle-surmounting simulation. Based on the Jacobian matrix and the initial parameter of the wheel-leg, the velocities of the actuators are fitted by the OLS （method of least squares）, which is the input of the simulation. After simulation, the dynamic characteristics of the actuators are obtained. The maximum power is less than the rated power of the actuator, which proves the feasibility of the intelligent obstacle-surmounting process. The test of surmounting the obstacle with 10 cm height is conducted for further validation. With the same structure, velocity and obstacle, the tests are conducted twice using the intelligent obstacle-surmounting method and the passive method respectively. The experimental result shows that the roll angle and trim angle by the intelligent obstacle-surmounting system are heavily decreased respectively from 4.5° and 2.5° to 0.75° and 0.4°compared with the passive method. Simulations and tests prove the validity of theoretical model and the effectiveness of intelligent obstacle-surmounting method.