Indoor Positioning And Path Planning For Mobile Robots Based On Swarm Intelligence Algorithms

With the development of robotic technology,robots are gradually becoming the key technical equipment to replace humans to complete heavy and highly repetitive work.As a core research component in robotics research,mobile robot navigation enables robots to perceive the environment through sensors and move automatically from the start point to the end point.Positioning and path planning are two key technologies in mobile robot high-accuracy navigation.Besides,optimizing the positioning scheme and navigation path of mobile robot can both reduce the working cost and improve the working efficiency.Therefore,this paper conducts the following research on indoor positioning technology and path planning method:(1)Study of the optimization method for sensor positioning network deployment.Positioning methods based on wireless sensor networks are one type of the common and effective positioning solutions for mobile robots.Since a reasonable sensor deployment can effectively improve positioning accuracy and reduce cost,an improved particle swarm optimization with virtual force guidance(IPSO-VF)is proposed and applied to sensor deployment optimization.First,an S-shaped inertia weight dynamic update strategy is proposed to better balance the global and local search capabilities of the algorithm.Second,a modified velocity update mechanism is designed by introducing a random learning exemplar and a two-stage adaptive learning factor update strategy to enable the algorithm to have greater diversity of solutions in the early stage and simultaneously ensure rapid convergence to the optimal solution in the later stage.Then,a global historical best position disturbance strategy is developed,which introduces different degrees of perturbations to search for high-quality deployment solutions near the current optimal solution.Meanwhile,the concept of virtual force is introduced to guide the sensors by using the virtual force between nearby sensors to prevent the sensor distribution from being too aggregated.The simulation experiment and positioning experiment show that the sensor network deployed by IPSO-VF achieves a coverage rate of 99.6%and centimeter-level positioning accuracy at all measurement points.(2)Study of the fusion positioning scheme based on Ultra Wide Band(UWB)and inertial navigation.Although the influence of static obstacles on positioning accuracy could be reduced by sensor deployment optimization,unknown dynamic obstacles in the indoor environment may still cause interference to the signal propagation of UWB.Therefore,this paper designs a fusion positioning system based on UWB and inertial navigation to improve the accuracy and working stability of the positioning scheme.The error state Kalman filter is introduced in this paper to estimate the error state of the navigation information and then use the error state to correct the system.The data of inertial measurement unit are used in the prediction session to obtain the predicted error state and the nominal state,while UWB data are utilised as observation to correct the predicted error state.The fused positioning results are acquired by combining the corrected error state with the nominal state.The results of the positioning experiment show that the average positioning error of the fusion system is reduced by 1.6 centimeters compared with that of the single UWB positioning system,indicating that the fused system possesses higher positioning accuracy.Moreover,in the extreme case of missing UWB data,the average positioning error of the fused system is still lower than 5 centimeters,which demonstrates that the fused positioning system has higher working stability and anti-interference capability.(3)Study of the robot path planning method.Efficient path planning can ensure that the robot reaches the end point from the starting point quickly for the next operation,which could improve the work efficiency.Therefore,this paper proposes an improved grey wolf optimizer algorithm(IGWO)for path planning problems.Firstly,a new update mechanism is designed to better balance the exploration and exploitation ability of the algorithm.Then,a dynamic local optimum solution escape strategy is proposed to improve the ability of the algorithm to jump out of the local optimum value.In addition,the individuals ranking low in the population will be repositioned to accelerate the convergence speed of the algorithm.Moreover,this paper adopts the cubic spline interpolation as smoothing method to obtain paths that better match robot dynamics.In order to verify the effectiveness of the proposed algorithm,simulation experiments are conducted in simple obstacle environment and complex irregular obstacle environment,and the results indicate that IGWO can plan shorter paths with 100%and 90%success rates in the two environments respectively,indicating that the proposed IGWO has better path planning capability and higher stability.In order to further verify the practicality of the path planning method,this paper uses a real robot platform and the proposed fusion positioning system to conduct navigation experiments.The results show that IGWO can help the robot avoid obstacles and reach the end point successfully in the real environment,confirming the feasibility of the proposed path planning method.