基于脉冲神经网络的智能小车的自动避障系统研究

Research on Automatic Obstacle Avoidance System of Smart Vehicle Based on Spike Neural Network

为了提高智能小车自动避障控制的精度,本文提出了一种基于脉冲神经网络(SNN)的避障控制模型。该神经网络采用泊松编码方式,使网络模型不仅具有更明确的生物意义,还具有强大的计算能力。本文针对全连接网络结构导致计算量大的问题提出了一种基于STDP规则的稀疏概率连接网络结构,通过实验验证,该结构有效地减少了网络的训练时间,同时还提高了算法的效率和准确率。采用多个超声波传感器探测障碍物的方位,信息经过脉冲神经网络处理后,实现了智能小车对障碍物的安全避障控制。考虑小车与障碍物和目标点的距离、角度以及小车的行驶速度等多种路况,该模型采用灰色关联分析法选取相似路况进行模拟验证。本文分别对基于SNN、支持向量机(SVM)和BP人工神经网络(BP-ANN)的控制模型进行测试和评估。实验结果表明:SNN避障控制模型相比于其他模型有更高的控制精度和适用性,为智能车辆的开发避障控制算法提供了一定的理论依据和应用价值。

An obstacle avoidance control model based on Spiking Neural Network (SNN) was proposed in this paper to improve the accuracy of the automatic obstacle avoidance control of the smart vehicle. This neural network uses Poisson coding, which not only has a clearer biological meaning, but also has powerful computing capabilities. This paper proposes a sparse probability connection network structure based on STDP rules for the problem of a large amount of calculation caused by the fully connected network structure. Through experimental verification, this structure effectively reduces the training time of the network and also improves the efficiency and accuracy of the algorithm. Multiple ultrasonic sensors are used to detect the orientation of obstacles. After the information is processed by the pulse neural network, the smart car realizes the safe obstacle avoidance control of the obstacles. Considering the road conditions such as the distance and angle between the trolley and the obstacle and the target point, as well as the traveling speed of the trolley, the model adopts the gray correlation analysis method to select similar road conditions for simulation verification. This paper tests and evaluates control models based on SNN, support vector machine (SVM), fuzzy neural network and BP artificial neural network (BP-ANN). The experimental results show that the SNN obstacle avoidance control model has higher control accuracy and applicability than other models, which provides a certain theoretical basis and application value for the development of obstacle avoidance control algorithms for intelligent vehicles.