By analysis on operation characteristics and environments of seabed miner

the key problem is the local minimum when traditional potential field method is applied on obstacle avoidance. An improved algorithm based on artificial potential field method was put forward. The potential field was redefined and the repulsive potential function was improved. The kinematics model of crawler miner was constructed. It was applied on realtime obstacle avoidance of miner in seabed unknown environment successfully. The simulation results show that the robot can skip from the local minimum and get to the goal smoothly. The improved algorithm is practical and feasible.The research has an instructive meaning for seabed mining and obstacle avoidance of underwater robot.