Nowadays we are aware of accelerated development of automotive software. Numerous of ADAS (Advanced Driver Assistance Systems) systems are being developed these days. One such system is the forward CAS (Collision Avoidance System). In order to implement such a system, this paper presents one solution for detecting an object located directly in front of the vehicle and estimating its distance. The solution is based on the use of camera and LIDAR (Light Detection and Ranging) sensor fusion. The camera was used for object detection and classification, while 3D data obtained from LIDAR sensor were used for distance estimation. In order to map the 3D data from the LIDAR to the 2D image space, a spatial calibration was used. The solution was developed as a prototype using the ROS (Robot Operating System) based Autoware open source platform. This platform is essentially a framework intended for the development and testing of automotive software. ROS as the framework on which the Autoware platform is based, provides a library for the Python and C++ programming languages, intended for creating new applications. For the reason that this is a prototype project, and it is popular for application in machine learning, we decided to use the Python programming language. The solution was tested inside the CARLA simulator, where the estimation of the obstacle distance obtained at the output of our algorithm was compared with the ground truth values obtained from the simulator itself. Measurements were performed under different weather conditions, where this algorithm showed satisfactory results, with real-time processing.
