Hexapod Robot Collaborates Uav For Special Environment Air Pollution Prediction

Since the Industrial Revolution of the Western Powers,a large number of industrial levels have developed rapidly,human behavior gradually had a great influence on nature,routine index of the atmosphere environment such as temperature,humidity,PM_2.5 and the concentration of sulfur and nitrogen oxides in air directly reflects a regional air quality and pollution level,especially in complex unstructured environment,such as nuclear site,high altitude mountains,deserts,wetlands and uneven road surface such as the gobi.The concentration of collection and state prediction of atmospheric pollutants in such areas have become a research hotspot in the field of environmental protection.Aiming at the complicated condition forecasting,real-time acquisition and the special environment air pollutants is proposed based on a six foot robots and unmanned aerial vehicle（uav）cooperative data collection and analysis system,the system combines the advantages of UAV’s high-performance,flexible action and wide acquisition range,and the adaptability of the hexapod robot’s unstructured road surface,multi-degree of freedom and high flexibility.Subsequently,the prediction model of XgBoost and random forest environmental pollution is established to predict and compare the performance of collected pollution index concentration.The specific research contents of this paper are as follows:（1）The walking and analysis system of hexapod robot is established.The hexapod robot has the functions of autonomous movement,sensing detection and environment recognition.By using the upper computer data processing platform,the collected air pollutant indicators are cleaned and pretreated,and the subsequent comprehensive information of the concentration of each indicator and environmental characteristics is used to predict the air pollution situation in this region.（2）Establishing the uav data acquisition system.Using both the advantages of flexibility,stability and can load the six rotor small unmanned aerial vehicle（uav）as the bearing platform,carrying embedded microcontroller controller and all kinds of environmental pollution indicators acquisition sensors,such as temperature and humidity sensor,PM_2.5 sensor such as sulfur,nitrogen oxide sensors,which can realize a region of atmospheric pollutants concentration sampling and data classification.（3）The real-time multi-aircraft wireless communication system of hexapod robot and UAV is built.The improved PSO algorithm is used to establish the information sharing channel of environmental fitness through the perception and information acquisition of the environmental adaptation of UAV and hexapod robot respectively,and to share the environmental fitness of each robot in real time,which makes the whole multi-aircraft system adjust to the optimal position and speed,and UAV acquisition.The concentration of air pollution index is transmitted to the upper machine system of ground hexapod robot through wireless communication system to ensure that UAV and hexapod robot can maintain the optimal traveling speed and path in complex unstructured environment.（4）After obtaining atmospheric pollutant index concentration and environmental characteristic factors,a data-driven prediction method based on Xgboost and random forest was designed.This method has great anti-interference ability when facing multi-dimensional feature vector and sparse feature input.It has the ability to accurately predict the regional pollution status after synthesizing various air pollution indexes.By comparing the predicted results of random forest and Xgboost algorithm.After comparing the actual test with the actual data,it is found that there is still room for improvement in prediction accuracy and accuracy,which is related to the feature engineering establishment and model parameter optimization,which will be improved in the future.