System Design And Realization Of Intelligent Building Inspection Robot

With the continuous advancement of modern city construction,people’s requirements for building safety are increasing.It is difficult to meet the needs of modern building security by simply relying on manual periodic inspections.When adopting the detection method of fixed position sensor,it has limitations due to the existence of detection blind spots,the complexity of networking,and the high installation cost.In response to the above problems,this article has developed an intelligent inspection robot for building security based on the needs of the field of modern building security.The robot uses Turtle Bot3 as the mobile platform.This article uses sensors,deep learning and intelligent voice technologies to design and implement the robot system with multiple functions.The functions mainly include map construction and automatic navigation and obstacle avoidance,video surveillance,environmental monitoring and intelligent voice recognition and Web background management,etc.,so as to realize the dynamic inspection of the building all-weather and without blind spots.The main work of this paper is as follows:(1)This paper adopts the Cartographer algorithm to realize the construction of the map of the inspection site and achieve a good mapping effect.Global path planning algorithm Dijkstra and A* are analyzed and compared by simulation,and A* algorithm with better performance is selected.Besides,Dynamic Window Approach algorithm is used for local planning.In the field test,the target function was successfully achieved.(2)The obtained video frames are sent to the background for real-time display.Visible light and infrared video surveillance are mainly used in video surveillance.Visible light monitoring realizes functions such as object detection and face recognition,and realizes the display and transmission of video results.Infrared video surveillance can extract temperature information in the video for instrument fault detection,and use Keras to build a convolutional neural network to support instrument recognition of infrared images.(3)This paper uses various sensors on board to collect environmental parameters,and realizes early warning of environmental abnormalities.The system not only can actively perform voice alarms,but also supports intelligent voice recognition to realize basic interactive functions such as voice inquiries.This system can be managed by the web interface and realize the visual operation of each data.Web interface realizes local and remote control of the robot pan-tilt and chassis,which solves the problem in the blind spot of monitoring.(4)This paper implements the web background management system interface of the inspection robot,which can remotely control the overall functions.The function includes the parameter threshold setting of each sensor and the voice broadcast of Chinese characters,etc.In addition,it realizes the visualization of various monitoring data and the loading of video monitoring screens,as well as the realization of docking with the building Internet of Things system.The functional modules of the intelligent inspection robot have successfully passed the field test.Its high-efficiency and all-round dynamic inspection is of great significance to improving the safety and intelligence of the building,and has broad practicability and development prospects.