Indoor Positioning And Navigation System Based On Computer Vision

Indoor positioning and navigation technology is currently a hot research technology,and with the increasing number of large buildings,this technology is becoming more and more important.Indoor positioning and navigation is different from the outdoor navigation technology commonly used by GPS or Beidou satellite navigation.Due to the blocking and refraction of the wall,the outdoor satellite signal will have a large deviation.In order to obtain more accurate positioning,indoor sensor positioning is needed..Path planning and navigation are equally important after positioning.This paper aims to establish a fast and accurate indoor positioning and navigation system.The positioning technology of this paper is based on the principle of radio frequency and ultrasonic ranging.It builds multiple radio frequency and ultrasonic positioning base points indoors,and measures the distance between the positioning base point and the moving target-smart car.The radio frequency positioning base point completes the rough positioning of the smart car.The ultrasonic positioning base point completes the precise positioning of the smart car.Finally,the indoor coordinate error of the smart car is not more than 5cm,which basically meets the positioning accuracy required by the system.The indoor path planning algorithm of this paper is based on the improved parameter adaptive ant colony algorithm.Ant colony algorithm is a bio-intelligence algorithm developed according to the principle of ant pathfinding.When searching for food,ants will leave pheromone on the path.As time accumulates,the pheromone concentration on the shortest path is higher than other paths.This allows subsequent ants to have a greater chance of selecting the path.In this paper,the main parameters of ant colony algorithm are improved.The problem of slow convergence is caused by the lack of pheromone in the early stage of ant colony algorithm.The weight parameters ? and ? of pheromone and heuristic information are improved,and two parameters are dynamically adjusted.In view of the high concentration of pheromone in the late iteration,the ant colony is easy to fall into the local optimal problem,and the pheromone evaporation coefficient is improved to make it a dynamic global adaptive parameter.Simulation experiments show that the improved ant colony algorithm has a faster convergence speed.The navigation principle designed in this paper is based on computer vision technology,which uses the popular convolutional neural network to identify images taken by smart cars.Based on the VGG network,a 16-layer convolutional neural network is designed.The input of the network is the picture taken by the smart car during the driving process.The output is the classification of the pictures.Different classification marks the different steering angles of the smart car.The collected images are marked by the location of the smart car on the planned route,and the network is trained to finally obtain the neural network model.Experiments show that the classification accuracy of the test set pictures is higher than VGG16.Choosing the right hardware module and software system,designing and combining the above three modules together constitute the indoor positioning and navigation system of the final design of this paper.