{"id":"arxiv/2604.24894","url":"https://arxiv.org/abs/2604.24894","tags":["cs.RO","cs.CV","cs.LG","eess.SY","math.OC"],"score":0,"title":"VISION-SLS: Safe Perception-Based Control from Learned Visual Representations via System Level Synthesis","domain":"cs.RO","source":"arxiv","authors":["Antoine P. Leeman","Shuyu Zhan","Melanie N. Zeilinger","Glen Chou"],"reviews":0,"subject":"cs.RO","summary":"We propose VISION-SLS, a method for nonlinear output-feedback control from high-resolution RGB images which provides robust constraint satisfaction guarantees under calibrated uncertainty bounds despite partial observability, sensor noise, and nonlinear dynamics. To enable scalability while retaining guarantees, we propose: (i) a learned low-dimensional observation map from pretrained visual features with state-dependent error bounds, and (ii) a causal affine time-varying output-feedback policy optimized via System Level Synthesis (SLS). We develop a scalable, novel solver for the resulting…","category":"research","comments":0,"disputed":false,"division":"cs","mentions":0,"discussed":0,"published":"2026-04-27","reactions":{}}