Search results

Perception has been identified as the main cause underlying most autonomous vehicle related accidents. As the key technology in perception, deep learning (DL) based computer vision models are generally considered to be black boxes due to poor interpretability. These have exacerbated user distrust and further forestalled their widespread deployment in practical usage. This paper aims to develop explainable DL models for autonomous driving by jointly predicting potential driving actions with corresponding explanations. The explainable DL models can not only boost user trust in autonomy but also serve as a diagnostic approach to identify any model deficiencies or limitations during the system development phase.

This paper proposes an explainable end-to-end autonomous driving system based on “Transformer,” a state-of-the-art self-attention (SA) based model. The model maps visual features from images collected by onboard cameras to guide potential driving actions with corresponding explanations, and aims to achieve soft attention over the image’s global features.

The results demonstrate the efficacy of the proposed model as it exhibits superior performance (in terms of correct prediction of actions and explanations) compared to the benchmark model by a significant margin with much lower computational cost on a public data set (BDD-OIA). From the ablation studies, the proposed SA module also outperforms other attention mechanisms in feature fusion and can generate meaningful representations for downstream prediction.

In the contexts of situational awareness and driver assistance, the proposed model can perform as a driving alarm system for both human-driven vehicles and autonomous vehicles because it is capable of quickly understanding/characterizing the environment and identifying any infeasible driving actions. In addition, the extra explanation head of the proposed model provides an extra channel for sanity checks to guarantee that the model learns the ideal causal relationships. This provision is critical in the development of autonomous systems.

Considerably diminished quality of life (QoL) is observed in patients with visual field defects after lesions affecting the visual pathway. But little is known to what extent vision-and health-related QoL impairments are associated with psychological distress. In 24 patients with chronic visual field defects (mean age=56.17±12.36) the National Eye Institute-visual functioning questionnaire (NEI-VFQ) for vision-related QoL, the Short Form Health Survey-36 (SF-36) for generic QoL and the revised Symptom-Checklist (SCL-90-R) were administered. Cases with clinically relevant SCL-90-R symptoms were defined. Demographic, QoL and visual field parameters were correlated with SCL-90-R scales. About 40% of the investigated patients met the criteria for the definition of psychiatric caseness. 8/12 NEI-VFQ scales correlated significantly with SCL-90-R phobic anxiety (r-range -0.41 to -0.64, P<0.05), 5/12 NEI-VFQ scales correlated with SCL-90-R interpersonal sensitivity (-0.43 to -0.50), and 3/12 with SCL-90-R depression (-0.51 to -0.57) and obsessive-compulsiveness (-0.41 to -0.43). In contrast, only 1/8 SF-36 scales correlated significantly with SCL-90-R depression, phobic anxiety and interpersonal sensitivity (-0.41 to -0.54). No substantial correlations were observed between visual field parameters and SCL-90-R scales. Significant correlations of SCL-90-R with NEI-VFQ but not with SF-36 suggest that self-rated psychological distress is the result of diminished vision-related QoL as a consequence of visual field loss. The extent of visual field loss itself did not influence the rating of psychological distress directly, since SCL-90-R symptoms were only reported when diminished vision-related QoL was present. Patients with reduced vision-related QoL due to persisting visual field defects should therefore be offered additional neuropsychological rehabilitation and supportive psychotherapeutic interventions even years after the lesion.