Systematic Derivation of Requirements for the Perception Task of Free Space Driving Assistance Function applied to Trucks

The larger size and expanded blind spots of heavy-duty trucks in comparison to passenger cars, create unique challenges for truck drivers navigating narrow roads, such as in urban scenarios. For this reason, the detection of free space around the vehicle is of critical importance, as it has the potential to save lives and reduce operating costs due to less maintenance and downtime. Despite the existence of numerous approaches to free space detection in the literature, few of these have been applied to the trucking sector, disregarding important aspects for these kinds of vehicles such as the altitude at which obstacles are located. This paper aims to present the initial results of our research, a "Not Free Space Warner", a driving assistance function intended for implementation in series trucks. A methodology is followed to define the characteristics that the perception component of this function shall fulfill. To this end, an analysis of the most critical accidents and common driving situations that truck drivers encounter is conducted, with a particular focus on the potential contribution of free space detection to assist the driver. By deriving and analyzing multiple scenarios from the use cases, the requirements to be met by the perception pipeline of function are defined. To validate these requirements, a Mercedes Actros equipped with multiple ground truth sensors, utilized as a research vehicle, is presented. Finally, the limitations and challenges associated with its implementation in the context of trucks are discussed.