Object Classification with 2D LiDAR Distance Data: A Machine Learning-Based Approach

LiDAR technology is widely used in areas such as object detection and classification. This study investigates whether object classification is possible using distance data from a 2D LiDAR simulation developed on Unity, and evaluates it using various machine learning methods. In the simulation environment, distance data for five different objects were collected using a LiDAR sensor, and feature extraction was performed to provide input for classification models. The performance of random forest, kNN, and artificial neural networks methods was compared. Experimental results showed that random forest achieved the highest performance with an accuracy rate of 97.47%, particularly providing nearly flawless classification for the sphere object. The findings demonstrate that successful object classification can be achieved using only 2D LiDAR distance data, and ensemble methods provide higher accuracy on such data. The study evaluates the potential of low-cost, less complex 2D LiDAR systems in object detection and presents important findings on when these systems can be effective.