Distributed Cooperative Localization Framework Via Consensus Optimization

Continuous and reliable flow of navigation information is a pressing need for safer and smarter transportation systems. Vehicles with various levels of autonomy are now penetrating the markets. However, legacy vehicles (LVs) constitute a large portion of vehicles still running to this date. Hence, schemes for leveraging the coexistence of all vehicle types are required. This work represents an attempt towards achieving this goal. The paper proposes a distributed framework to leverage the computational capacity and advanced sensing capabilities of sensor-rich vehicles (SRVs) to enhance the localization accuracy of all participating vehicles. Neighboring SRVs and LVs form an ad-hoc network to exchange information through V2V communication, and SRVs distributively solve a consensus localization optimization problem. Thus, the system eliminates the need for a centralized entity that aggregates the data from all participating vehicles and solves the localization problem. The numerical results show that the proposed system achieves the same performance levels of the centralized methods. It is also shown that all participating vehicles including SRVs and LVs can benefit from this cooperation. For example, localization errors are reduced by at least 62:8% and 11:9% for LVs and SRVs in one scenario, respectively.