Trajectory estimation and prediction in a convoy of vehicles

Abstract

The last-mile delivery problem is increasingly a challenge to local authorities and government actors. 25% of the road congestion is due to delivery vehicles which also generate air and noise pollution. This is a waste of time and energy that must be avoided for economic and environmental reasons. Expleo Group, responds to this problem by developing a solution that revolves around a vehicle convoy system, consisting of a leader vehicle and several automated tracking vehicles that are dropped off and picked up in dedicated delivery zones. The solution proposed by Expleo is based on a convoy bound by a dematerialized "Truck Platooning" hook which relies on a combination of target tracking by camera and Vehicle to Vehicle (V2V) communication, unlike what already exists now (use of radars, LIDARs, and other expansive technologies). This project, "Smart Delivery" is a use case of the global case "Smart Platooning" whose objective is the development of vehicle platoons in general. This report includes a series of works carried out around the platooning (vehicle convoy driving) whose objective is to estimate the trajectory of each vehicle in the convoy using the V2V communication, which will provide crucial information to the proper functioning of the system. The approach developed in this document depends heavily on the communication between the following vehicles and the leading vehicle, which has a very important role in the formation and guidance of the convoy. This convoy is based on information from onboard sensors such as odometers, GPS, cameras, etc., and the exchange of information between the vehicles of the platoon via V2V communication (see Figure 1). As the vehicles are communicating through V2V communication, they can anticipate the movement of each vehicle and behave smoothly avoiding the concertina effect. The estimation of the position of each vehicle must be as accurate and robust as possible since it will be part of the inputs of the lateral and longitudinal controllers that are required to respectively track the leader’s trajectory and keep a desired spacing between them. This spacing varies with the mass and the velocity of each vehicle. To accomplish this task properly, the mission is divided into 3 parts: the estimation and localization part, to know at each instant where are the vehicles, the control part, to make them follow properly a given trajectory and the fault detections part, to make the system more robust and able to stop in case of dysfunction. Only the first one will be detailed in this document. My colleagues intern worked on the other parts