Nowadays,the population in China is showing an aging trend,so the problem of accompanying the elderly turns into increasingly prominent;In addition,the lack of care for only children has also become a phenomenon in most families.With the development of robot technology,the escort robot have gradually changed the family lifestyle,which will effectively solve the care problem of the elderly and children.In order to meet these escort needs,this paper designs a ROS-based escort robot with the ability to SLAM and navigation faster and more accurately,while its voice interaction is more intelligent.The main work of this paper is as follows.(1)The overall design of the system.The overall design is explained from the system frame design,robot model design,hardware platform design and ROS architecture analysis.The system framework reflects the overall design idea,and divides the system into control layer,functional layer and hardware layer.Robot model design includes the design of robot system model,sensor observation model,and robot URDF model.The hardware platform design mainly completes the device selection and the overall construction robot.The system architecture,communication mechanism and common components are used to analyze the ROS architecture.(2)SLAM and navigation design.SLAM design introduces improved Gmapping algorithm for RBPF-SLAM,which reduces the number of particles and solves the problem of particle degradation through improve proposed distribution and adaptive resampling.Navigation design involves positioning and path planning.The former uses an Adaptive Monte Carlo Localization algorithm to improve the response speed and ensure the positioning accuracy.The latter includes global path planning and local path planning.Global path planning utilizes smoothed A* algorithm to minimize redundant inflection points and obtain the best path.Local path planning adopts DWA algorithm to calculate the local optimal path when running between adjacent nodes in the form of an evaluation function.(3)Voice interaction design.After completing the speech interaction design framework,the HMM-based speech recognition system and waveform splicing-based speech synthesis system are modeled respectively.During the realization analysis of voice interaction,the tasks such as the speech dialogue,the voice control,and expression transformation are completed and the education,learning,interactive entertainment,and emotional expression features of escort robot are achieved.In performance testing and analysis,the SLAM and navigation tests are accomplished with two methods including URDF model and real robot.The test results show that the speed and accuracy of robot in building maps and navigation are significantly improved compared with the algorithm before improvement.The voice interaction test is implemented by examining the speech dialogue,the voice control,and expression transformation,etc.The overall test results of the system verify the practicability and effectiveness of the designed home escort robot. |