Automatic Navigation Technology Of Inspection Robot For Natural Gas Station

The research and development of inspection robot for natural gas station can greatly decrease the maintenance cost and increase the safety performance.Based on the condition of the specific natural gas station,the grade for the explosive-proof reaches ECII,the narrowest region width is 0.9m,and the location precision of GPS is 8cm.All this means that the inspection robot must be equipped with fine load property,compact mechanical structure and high precision location system.Therefore,this research chooses wheeled skid-steer mobile robots(WSMR)as basic structure,develops a drive system(LDS)based on CAN bus,builds GPS/INS integrated navigation system and the upper computer control system.For better control,kinematics and dynamics of WSMR are intensively studied in this research: the kinematic model is built with fully consideration of axial velocity and slip rate during the movement of WSMR.the relation of slip displacement and shear stress of wheels and ground are studied under nonlinear constraint,further integration is made to get the stress condition of each wheel.experiments in different ground conditions(brick floor which is similar to the natural gas station and marble floor)are conducted to get the loss coefficient of turn in low linear velocity rotary motion,and is 1.89 in brick floor and 1.69 in marble floor,and loss coefficient of turn is positively related to coefficient of sliding friction.The simulation of dynamical model presents the stress condition of car turning of each wheel,it shows that the motor torque of front wheel is larger than the back wheel's,and the motor torque of the outside wheel is larger than the one of the inside wheel.The design of integrated navigation and location system is an important part of the realization of the inspection robot.In this research,quaternion is chosen as pose description method,complementary filter is chosen to correct data from inertial sensor by considering the small variation of the robot,further Kalman filter is performed to process the data of the location and the navigation GPS output and the data corrected by complementary filter.The result shows that pose data from complementary filter can depict the real pose better,and heading information after Kalman filter fluctuate smaller in steady motion and is not disturbed by random errors caused by gyroscope.According to the basic inspection requirement of the robot,design real time motion path planning algorithm based on the shortest path is proposed.Halton sequence is chosen to sample the map data with considering the kinematics performance of the robot,by using dynamic programming algorithm,the shortest path is achieved.PID control algorithm is chosen for preplanning by considering the smallest radius of turning circle,and based on pre-plan and optimized reference points from dynamic trajectory,the path of the robot is optimized.Thus,the robot motion control plan is finished.The conclusion of big size map simulation shows that algorithm proposed in this research has ideal effect in the turning radius stability and obstacle avoidance performance.Also,in the same mini map,compared with the SBMPC,the proposed algorithm has better path distance and real-time performance.And compared with the general shortest path algorithm,the turning radius is more stable and energy consumption is less in the proposed algorithm.This research finished a complete development of the basic navigation of inspection robot for natural gas station.