An Intelligent Navigation Robot Based On RSSI Localization

In connection with the problems existing in legacy RSSI-based localization technology such as high cost,low accuracy,unsuitable with environment complexity or variability,we designed a wheel robot that combined with LIDAR SLAM and RSSI based localization,implementing functions which mainly are creating map of place where robot is at when robot is in motion and navigating from current place to designated goal automatically.Mini system Soarsea V200 is adopted as central controlling computer of the robot.Microcontroller STM32F103 RC is introduced into the design of the robot purposed to control other hardware modules including inertia measurement unit MPU6050,motor driver A4960,and direct current reduced motor JGB37-520,etc.except RPLIDAR A1M8 and act like a middleware between central controlling computer and outer circuits.Work principles of mobile robot system are described in the first place.Functional and performance requirements that an arbitrary mobile robot system should fulfill are analyzed.The overall architecture design of our mobile robot based on these principles and requirements is proposed.Hardware system solution of our mobile robot is brought forward to meet the designing requirement,including appearance,hardware architecture and chosen components with interaction program module that enable encapsulated communication between controller and other components.Components of our robot hardware system consist controller host computer,microcontroller,inertial measurement unit,motor with driver chip and LIDAR.Software modules that implement main functions of our robot are selected under Robot Operating System framework,involving gmapping SLAM,hector SLAM,global path planner,local path planner,trajectory planner and map service package.RSSI-based localization module is designed cooperating with ROS framework.An improved RSSI distance estimating model is proposed in order to promote accuracy of result and reduce error,by taking signal properties like electromagnetic signal reflection,refraction,etc.and environmental condition in account and representing their influence on signal measurement as different parameters of our new model.A self-adaptive method which adjusts parameters of algorithm accordingly to surrounding condition change is proposed,makes accuracy of distancing result better.Triangulation is introduced in data processing and data fitting when constructing data noise filter,which avoided limitations of simple mean or medium filtering method.By running simulation it is found that accuracy and stability of our model is better than the original logarithmic model,and overall accuracy of dynamic algorithm is better than static algorithm.The average error of self-adaptive method is nearly30% below with the average error of legacy model.Signal strength distribution over environment shows significant characteristics that chaos system has according to what simulation result expressed.