{"paper":{"title":"Still Camouflage, Moving Illusion: View-Induced Trajectory Manipulation in Autonomous Driving","license":"http://creativecommons.org/licenses/by/4.0/","headline":"A static camouflage on one vehicle can make passing autonomous cars see false cut-in trajectories and brake hard.","cross_cats":["cs.CV"],"primary_cat":"cs.CR","authors_text":"Feng Liu, Haotang Li, Huashan Chen, Kebin Peng, Qingzhao Zhang, Sen He, Shuo Ju, Wanqian Zhang, Xuheng Wang","submitted_at":"2026-05-12T20:47:55Z","abstract_excerpt":"Existing physical adversarial attacks on vision-based autonomous driving induce time-evolving perception errors, including biased object tracking or trajectory prediction, through (i) sophisticated physical patch inducing detection box drift when entering the view distance, or (ii) dynamically changing patches that cause different perception errors at different time. In both cases, viewing-angle variation is treated as a challenge, requiring adversarial patches to remain effective across frames under varying views, leading to complex multi-view optimization. In contrast, we show that viewing-a"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"We demonstrate the novel attack on nuScenes dataset, showing the effectiveness with an end-to-end success rate of up to 87.5%, measured by hard-braking events, and robustness across different scene backgrounds, victim vehicle speeds, and perception models.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The assumption that view-dependent feature drift from a static camouflage will reliably propagate through the full perception-to-planning pipeline in real-world lighting, sensor noise, and diverse perception models beyond the tested ones.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Static adversarial camouflage exploits natural view-angle changes during relative motion to induce consistent feature drift in AV perception, leading to incorrect trajectory predictions and unnecessary braking.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"A static camouflage on one vehicle can make passing autonomous cars see false cut-in trajectories and brake hard.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"ba10fbf18e28971eecb5b99792368af436a217a32749680f3aa12148246c1c81"},"source":{"id":"2605.12743","kind":"arxiv","version":1},"verdict":{"id":"018ba712-aeca-4a5e-acea-71b91d39fe00","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-14T19:39:58.308022Z","strongest_claim":"We demonstrate the novel attack on nuScenes dataset, showing the effectiveness with an end-to-end success rate of up to 87.5%, measured by hard-braking events, and robustness across different scene backgrounds, victim vehicle speeds, and perception models.","one_line_summary":"Static adversarial camouflage exploits natural view-angle changes during relative motion to induce consistent feature drift in AV perception, leading to incorrect trajectory predictions and unnecessary braking.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The assumption that view-dependent feature drift from a static camouflage will reliably propagate through the full perception-to-planning pipeline in real-world lighting, sensor noise, and diverse perception models beyond the tested ones.","pith_extraction_headline":"A static camouflage on one vehicle can make passing autonomous cars see false cut-in trajectories and brake hard."},"references":{"count":56,"sample":[{"doi":"","year":2022,"title":"Baidu Apollo.https://apollo.baidu.com/, 2022","work_id":"b58512f6-f4c4-43aa-bdda-ef1f1c45916c","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2026,"title":"Nvidia-alpamayo.https://www.nvidia.cn/solutions/autonomous-vehicles/alpamayo/, 2026","work_id":"3654562c-8ca7-4d17-91c9-ab1d9ec6f981","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2026,"title":"Tesla Full Self-Driving.https://www.tesla.com/fsd, 2026","work_id":"ba00db71-d441-4930-ab73-f4ee60e835b0","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2018,"title":"Synthesizing robust adversarial examples","work_id":"6afc6e4d-fc2e-41e7-b37e-9254b6df4349","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2022,"title":"Transfusion: Robust lidar-camera fusion for 3d object detection with transformers","work_id":"93a581a0-8c32-46a8-828a-ab06a0a74b4c","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":56,"snapshot_sha256":"57edad913fc65e1aaf3f7f5220a298aec30d489e376d0c6830e3a85975449ef3","internal_anchors":2},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}