UAP-BEV: Uncertainty Aware Planning in Bird's Eye View representations

Vikrant Dewangan ¹ Basant Sharma ² Sarthak Sharma ¹ Tushar Choudhary¹ Aakash Aanegola¹, Arun Kumar Singh², K. Madhava Krishna ¹

(1- Robotics Research Center, IIIT Hyderabad, 2- University of Tartu, Estonia)

👉 TODO

• Code Release
• Add simulation scenarios
• Improve README

Literature review is present in the literature-review branch.

Abstract

Motion planning over cost maps generated via Birds Eye View (BEV) segmentation has emerged as a prominent approach in autonomous driving. However, the current approaches have two critical gaps. First, the optimization process is simplistic and involves just evaluating a fixed set of trajectories over the cost map. The trajectory samples are not adapted based on their associated cost values. Second, the existing cost maps do not account for the uncertainty in the cost maps that can arise due to noise in RGB images, BEV annotations etc. As a result, these approaches can struggle in challenging scenarios where there is abrupt cut-in, stopping, overtaking, merging, etc from the neighbouring vehicles.

In this paper, we propose UAP-BEV, a novel approach that models the noise in Spatio-Temporal BEV predictions to create an uncertainty-aware occupancy grid map. Using queries of the distance to the closest occupied cell, we obtain a sample estimate of the collision probability of the ego-vehicle. Subsequently, our approach uses gradient-free sampling-based optimization to compute low-cost trajectories over the cost map. Importantly, the sampling distribution is adapted based on the optimal cost values of the sampled trajectories. By explicitly modelling probabilistic collision avoidance in the BEV space, our approach is able to outperform the cost-map based baselines in collision avoidance, route completion, time to completion, and smoothness.

Sources of Uncertainity

• Noise in RGB Images

• Noise in Intrinsics, error in GPS.

• Noisy BEV annotations

Dealing with Uncertainty - Parametric Assumption

• Assumes shape of underlying distribution
• Can lead to conservative behaviour
• Does not handle Noisy Annotations

We propose …. UAP-BEV

Part 1: Augmenting Uncertainty into Distance Estimates

Part 2: Efficient Batch Optimization with Constrained Projection

Results

Longitudinal Barrier Ablation

CEM Analysis

a. The histogram of constraint violation functions for each set in the elite set. At the start, multiple samples are spread out on non-zero values, complete convergence is achieved and at Iteration 10, all samples have values 0.

b. Convergence of normalized costs with trace of covariance, signifying the samples are pushed towards good regions.