LUCID: An Agentic AI Framework on Digital-Twin in the Loop for QoS-Guaranteeing Robotic Control
Authors: Hyeonsu Lyu, Minwoo Kim, Sehyun Ryu, Hyun Jong Yang
Organizations: Department of Electrical and Computer Engineering and the Institute of New Media and Communications, Seoul National University, Seoul, Republic of Korea · Pohang University of Science and Technology (POSTECH), Pohang, Republic of Korea
Cloud robotics relies on the timely uplink of high-volume sensing streams, yet dynamic environments continually shift the feasible combinations of trajectories, active-robot count, and per-robot QoS. Because existing approaches formulate trajectory planning (TP) and radio resource management (RRM) as a single fixed optimization problem, they cannot reconfigure these coupled decisions as conditions evolve, resulting in transient QoS violations. However, evolving operator intents change which quantities-such as the active-robot count and per-robot QoS-are fixed, optimized, or relaxed. Furthermore, the computational cost of evaluating trajectory-dependent wireless conflicts has made it difficult to build large-scale Digital-Twin-in-the-Loop (DITL) testbeds responsive enough for such dynamic orchestration. We present LUCID, an LLM-agent--orchestrated, uplink-aware cloud-robotics pipeline that moves TP--RRM from solving a fixed formulation to dynamically orchestrating optimization problem schemas within a DITL environment. Driven by the operator's high-level intent, LUCID treats the TP--RRM formulation as a bounded template whose variables, objectives, and constraints are dynamically configured, while SimBridge enables repeated ray-tracing evaluation by converting large-scale robotics scenes into wireless-ready DTs. By integrating collision-free path planning with a spectral-radius RRM validator, LUCID identifies wireless bottlenecks and restructures the problem schema on the fly to efficiently find the verified feasible state. Experiments confirm that LUCID robustly adapts to changing intents, active-robot counts, and scenes, while a multimodal surrogate model, FastConfigNet, reduces planning latency.
Figures & tables
Fig. 2: Two possible conflict types in warehouse environments.
Fig. 3: Uplink SINR map of two UEs over three steps. The heatmap shows the SINR of the red UE when the orange UEs act as interferers.
Fig. 4: DITL workflow for intent-conditioned TP–RRM evaluation.
Fig. 5: RRM coupling matrix for feasibility and conflict diagnosis.
Fig. 6: Per-slot RRM feasibility evaluation and conflict localization.
Fig. 7: Collision-free navigation-graph construction from the DT scene.
Fig. 8: Multimodal encoding architecture of FastConfigNet.
Fig. 9: Scene adaptation pipeline from a source DT to a wireless-ready scene.
Fig. 10: VLM-assisted mapping from asset metadata and a thumbnail to an ITU electromagnetic material identifier.
Fig. 11: Material-preserving mesh simplification in SimBridge.
Fig. 12: Mesh-reduction trade-off on the controlled small-warehouse scene. Left : USD and Mitsuba XML–PLY size. Right : path-gain RMSE relative to the 100% mesh and scene-load-plus-solver time.
Fig. 13: Performance of LUCID under changes in operator intent, active AMR count, and physical environment.
Fig. 14: End-to-end comparison of LUCID and Separated TP/RRM .
Department of Computing Science, Ume˚a University, 90187 Ume˚a, Sweden · School of Electrical Engineering and Computer Science, KTH Royal Institute of Technology, 10044 Stockholm, Sweden
Graduate School of Information Science and Technology, The University of Osaka, Osaka, 565-0871, Japan · D3 Center, The University of Osaka, Osaka, 560-0043, Japan