|
In summary, we carry out a research on a vehicle anti-collision warning system based on road condition detection. This system colletcts real time features of the road to determine the road types, cement of SMA, and the surface condition, dry, wet or icy.Then,it calculates road adhesion coefficient related to the size of braking distance based on the collected data.Besides,we establish a model to measure braking distances. Finally, through the Matlab simulation, we construct a collision warning system based on the Fuzzy theory.The method to detect road types is by following subsequent steps:a) Analyzes different road surface characteristics and the wheel vibration excitation source b) Establishes a wheel vibration model, the wheel vibration spectrum is considered as the feature space of the feature classification c) Constructs a neural network classifier with the wheel frequency vibration spectrum eigenvectors as the input to identify the features of cement and the SMA features of bitumen.The method to detect road conditions is also described in this paper. Road temperature depends on road condition (dry, wet and icy) and solar radiation, and there is the nonlinear causality among them, thus, road condition can be identified indirectly by road temperature and solar radiation with BP neural network.Furthermore, with the establishment of braking distance model, we study the design of anti-collision warning system based on road detection, analyze the braking process, discuss the important parameters of the model, and investigate into the parameter settings of the anti-collision system based on the Fuzzy theory. Due to a variety of complex variables, considering the great randomness and fuzziness, road type, road condition, weather, reaction speed of the driver are considered as the fuzzy subset. The maximum velocity is determined by running a simulation of the subset, then, braking distance prediction and risk early warning is realized.At the end, we bring out the design of a vehicle anti-collision warning system based on road condition detection by comparison between the simulated results and the on-site test. We use RBF neural network and Single-band reconstruction algorithm to analyze the collected wheel vibration acceleration to identify road type with90%accuracy. We establish a temperature monitoring system to collect road surface temperature, air temperature and humidity and transmit them through wired or wireless channels to the BP network to determine the road condition. Finally, we use the road type and the road condition to build up an anti-collision warning system. It uses the Fuzzy theory to calculate the safe distance and sets out precautions to avoid collisions.
|