Deconfliction of Motion Paths With Traffic Inspired Rules

Abstract

In this paper we investigate how to resolve conflicting motions for mixed robot-robot and human-robot multiagent systems. This work is motivated by atypical driving conditions, such as parking lots, where driving rules are not as strictly enforced as on standard roads. As a result, navigation algorithms should take into account the human drivers' behaviors, especially if they prove to be in conflict with the common rules of the road. In this work we make use of safety barrier certificates with a direction bias to deconflict agents' behaviour in a near-to-collision scenario, in compliance with local traffic rules. We also propose a tool to identify the driving direction bias-both for human and autonomous agents.

Probabilistic Safety-Assured Adaptive Merging Control for Autonomous Vehicles

“…R s is the minimum allowed distance between two vehicles to avoid collision for safety. Different from most existing CBF work with deterministic perfect model information [12], [14], the stochastic model in Eq. 6 leads to infinite support of ḣs em and hence we consider the chance-constrained optimization problem to accommodate uncertainty with η ∈ (0, 1) as the desired confidence of probabilistic safety.…”

“…However, the forward-invariant property relies on the solution feasibility, i.e., as long as a control solution satisfying the CBF constraints can be found at each time step, then the safety for future time steps can always be guaranteed. In reality, we may not always find such a solution and alternative solutions include switching to a full braking mode [12], [14]. However, this could lead to serious consequences in autonomous driving scenarios, e.g., when there is another vehicle behind the ego vehicle, or it is too late for the AVs to stop.…”

Multi-Robot Collision Avoidance under Uncertainty with Probabilistic Safety Barrier Certificates

“…The constrained control space specified by PrSBC in ( 23) and ( 24) ensures the forward invariance of probabilistic safety in (13). Hence, we can reformulate the original QP problem in (6) with the PrSBC constraints. The probabilistic safety controller can thus obtained by minimally modifying the original controller u * that the system wished to execute.…”

“…In this case the robots will simply decelerate to zero velocities to ensure safety, which may cause the deadlock preventing the robots from achieving the goals. Some deconfliction policies for deterministic SBC can thereby be employed, such as the one suggested in [6]. Readers are referred to [6] for detailed solutions.…”

“…Some deconfliction policies for deterministic SBC can thereby be employed, such as the one suggested in [6]. Readers are referred to [6] for detailed solutions.…”