Research On Key Techniques Of Obstacle Avoidance Path Planning Based On Baxter Robot

As modem manufacturing gradually becomes more intelligent and automated,robots are increasingly used in various industrial manufacturing fields.Considering that the robot working environment is complex and varied,there are various obstacles in the environment,which requires the robot to effectively avoid obstacles and complete specified operations.At the same time,as much as possible to ensure that this barrier-free path is close to the global optimum or sub-optimal,the robot moves smoothly along the path.This paper takes a 7-DOF robotic arm of Baxter robot as an example to study the key technologies of obstacle avoidance path planning under the known conditions of obstacles in the environment.Including Baxter robot kinematics,collision detection technology and obstacle avoidance path planning algorithm.The two arms of Baxter have a redundant degree of freedom which cause innumerable possibilities for inverse solutions.The same pose at the end of the arm can be reversed to obtain innumerable arrays of joint angles.In order to select a specific optimization solution from these arrays,this paper performs an optimization solution based on the gradient projection method,and then uses the Newton iteration and the algebraic inverse solution error compensation method to reduce the error caused by the gradient projection method.Finally,a set of joint angles that meet both accuracy and optimization requirements are obtained.Collision detection is the basis of the robot's obstacle avoidance motion planning field,and it is required to return the collision information of the current moment and the next moment.This article simplifies the link rod of Baxter manipulators by using a cylinder instead of a complete link model to do the intersection test.At the same time,the obstacle model is decomposed with the bounding box tree.Cross-testing between cylinder and the bounding volume of obstacle can give a final collision result,which not only speeds up the collision detection speed,but also reduces the difficulty of calculating the distance between the connecting rod and the obstacle without collision.In order to better plan an obstacle-free path,this paper proposes a rapidly-exploring random tree path planning algorithm that incorporates an adaptive gravitational potential field.Based on the RRT algorithm,the new algorithm uses the gravitational potential field of target point to guide the growth direction of the random tree,and adapts the gravity coefficient to change the potential field size so that the random tree can grow a barrier-free path better and faster.The new algorithm also further optimizes the planned path to make the path shorter and closer to the global optimum or sub-optimal.Finally,the MATLAB and MoveIt! tools are used to simulate the algorithm in 2D and 3D space.Furthermore,performing obstacle avoidance experiments on the Baxter robot and analyzes the experimental data.It is verified that the obstacle avoidance path planning algorithm in this paper takes less time to plan a collision-free,and the planned path not only has a good repeatability,but also close to the global optimal path.