Timed Rule-Based Supervision of an End-to-End Autonomous Parking Policy

Summary

We study whether a manually specified runtime supervisor can correct recurring failures of an existing end-to-end parking policy in a fixed CARLA parking lot.

Abstract Summary

We study whether a manually specified runtime supervisor can correct recurring failures of an existing end-to-end parking policy in a fixed CARLA parking lot. The vision-based Transformer architecture is inherited from Yang et al.; our contribution is a timed, rule-based Parametric Safety Shield (PSS) applied to its control outputs. The PSS uses hand-calibrated speed, position, and duration thresholds to intervene in observed failure modes, including boundary exits, delayed braking, and stalled or oscillatory control. In the reported closed-loop evaluation, 16 held-out target slots and six initial poses are each evaluated in four rounds (384 attempts per configuration). Target success increases from 327/384 (85.16%) for the retrained policy to 375/384 (97.66%) with the PSS; mean position and orientation errors among successful attempts are 0.21m and 0.33 degrees. These results show an improvement within this simulator setup. The repeated attempts share one map, vehicle, and sensor configuration, and the PSS uses simulator world coordinates; thus the results do not establish generalization to other lots or real vehicles, or a formal safety guarantee.

Key Points

We study whether a manually specified runtime supervisor can correct recurring failures of an existing end-to-end parking policy in a fixed CARLA parking lot.
The vision-based Transformer architecture is inherited from Yang et al.; our contribution is a timed, rule-based Parametric Safety Shield (PSS) applied to its control outputs.
The PSS uses hand-calibrated speed, position, and duration thresholds to intervene in observed failure modes, including boundary exits, delayed braking, and stalled or oscillatory control.
In the reported closed-loop evaluation, 16 held-out target slots and six initial poses are each evaluated in four rounds (384 attempts per configuration).
Target success increases from 327/384 (85.16%) for the retrained policy to 375/384 (97.66%) with the PSS; mean position and orientation errors among successful attempts are 0.21m and 0.33 degrees.

Method

We retain the network architecture of Yang et al. [ 2 ] as the nominal controller. The new component in this paper is the post-inference PSS. Figure 1 shows the complete signal path and separates inherited and added components.

Cameras, vehicle state, target slot Inherited E2E Parking Transformer Proposed timed PSS guards and overrides Vehicle in CARLA executed command next observation and simulator position/velocity for PSS Fig. 1: Complete closed-loop architecture.

Method — Timed Rule-Based Supervision of an End-to-End Autonomous Parking Policy — Fig. 3: Topology of the PSS structured as a finite-state supervisor.

Experiments

The network architecture keeps the same as the original one and hyperparameters are slightly different. The network was trained on 4 NVIDIA RTX 4090 (total VRAM: 96 GB) with a batch size of 16. Adam optimizer is utilized with an initial learning rate of $7.5\times 10^{-5}$ and a weight decay of $1\times 10^{-4}$, while the beta values are 0.9 and 0.999 as default.

The training dataset, Gen 2B of the E2E Parking Dataset [ 1 ] , comprises 248 training and 104 validation trajectories. As illustrated in Fig.

Experiments — Timed Rule-Based Supervision of an End-to-End Autonomous Parking Policy — Fig. 4: Top view of the parking lot in CARLA.

Sources and Demonstrations

Figures are reproduced from Gao et al. See the full paper for experimental details and the project page, when the paper links one, for demonstrations and videos.