A Collaborative RTK Approach to Precise Positioning for Vehicle Swarms in Urban Scenarios

Location information is fundamental in nowadays society and key for prospective driverless vehicles and a plethora of safety-critical applications. Global Navigation Satellite Systems (GNSS) constitute the main information supplier for outdoor positioning, with worldwide all-weather availability. While the use of GNSS carrier phase observations leads to precise location estimates, its performance can be easily jeopardized in urban scenarios, where satellite availability may be limited or observations may be corrupted by harsh propagation conditions. The satellite shortage is especially relevant for Real Time Kinematic (RTK), whose capability to estimate a precise positioning solution rapidly decays with weak observation models. To address this limitation, this article introduces the concept of collaborative RTK (C-RTK), an approach to precise positioning using swarms of vehicles, where a set of users participate in the vehicle network. The idea is that users with good satellite visibility assist users that evolve in constrained environments. This work introduces the C-RTK functional model, an estimation solution and associated performance bounds. Illustrative Monte Carlo simulation results are provided, which highlight that, by exploiting the cross-correlation terms present among the users' observations, C-RTK improves their positioning their of accuracy and availability.