Ground classification based on optimal random forest model

Today mobile robots are widely used in medical, rescue, and service industries, and are gradually replacing labor costs and playing an increasingly important role in various industries. The correct recognition of ground types can help robots adapt to different environments and help them reach the target area efficiently and quickly. Random forest model is a classifier model with widespread applications and is often used to solve the ground classification problem of mobile robots. In response to the problem that the traditional random forest algorithm has many initial parameters and lacks a definite formula to obtain the optimal combination of parameters, which leads to the low accuracy of the traditional random forest classifier model. In order to extend the ground recognition method and improve the accuracy of the random forest model, this paper proposes a random forest classifier model optimized by using genetic algorithm to classify ground types. The experimental results show that the effect achieved by the optimized random forest model, in terms of accuracy, has a significant improvement in accuracy compared with the traditional algorithm, reaching a recognition accuracy of 93%.