Indoor Integrated Positioning System For Mobile Robot Based On Improved SINS

With the expansion and development of the city,the application scenarios of mobile robots are more and more extensive.In the indoor environment,accurate positioning information is premise of autonomous mobile robots.At present,multi-sensor fusion is mainstream scheme of indoor positioning.Because inertial navigation equipment is not easy to be interfered by the external environment and can operate independently,it makes the inertial navigation equipment become common data source of multi-sensor indoor positioning algorithm.In this paper,a set of integrated positioning system is established by using SINS,dead reckoning and Kalman optimal linear smoothing.Its performance is reliable and can meet the requirements of indoor autonomous positioning of mobile robot.Firstly,an improved SINS algorithm is proposed.The traditional strapdown inertial navigation algorithm only uses multi subsample algorithm to compensate non-communicative error when robot moves.The multi subsample algorithm is approximately derived from the equivalent rotation vector equation(Bortz equation),but it has principle error in the algorithm.In order to further improve the accuracy,the incremental data of gyroscope is quasi synthesized into a polynomial form,which is substituted into differential equation of attitude array.Then the Taylor expansion of differential equation is carried out to obtain attitude information without non-communicative error,so as to reduce attitude error that has the greatest impact on positioning accuracy of the system.Then,the three subsample algorithm is used to optimize compensation terms of rotation effect and sculling effect in velocity differential equation.Velocity information is calculated by combining attitude information and velocity differential equation.Finally,the position information of robot is obtained by integrating velocity information,so as to derive an improved strapdown inertial navigation algorithm with higher accuracyThen,the error equation of dead reckoning system is analyzed.Combined with the characteristics of strapdown inertial navigation system and dead reckoning system,the integrated localization system with strapdown inertial navigation system and dead reckoning system is established based on Kalman optimal linear smoothing.Kalman optimal linear smoothing uses a group of observations in a period of time to get the optimal estimation of error at each time.Error is fed back to the results calculated by the modified algorithm in two subsystems to improve the positioning accuracy of integrated localization system.Finally,according to the requirements of integrated localization system data source,the hardware platform of mobile robot based on STM32 is built.The data of inertial navigation equipment and wheel odometer can be transmitted to upper computer in real time and saved by upper computer through Bluetooth.Besides the location data can be analyzed and calculated by integrated localization algorithm.Two sets of typical trajectory positioning experiments are designed for indoor positioning of mobile robots.The experimental results show that the absolute error RMS values of integrated localization system based on traditional SINS and improved SINS are 8.81cm and 6.89cm respectively,so the integrated localization system based on improved SINS has higher positioning accuracy.