Learned physics-based models provide structure for robot control, but uncertain payloads and unmodeled dynamics limit their predictive accuracy. Context-aware Deep Lagrangian Networks enable adaptation through a latent context inferred from measurement histories while retaining Lagrangian structure for each fixed context. However, this structure alone does not constrain sensitivity to measurement disturbances. We introduce BoBDeLaN, which combines a contracting, gain-bounded Recurrent Equilibrium Network context encoder with stable linear filtering to shape disturbance sensitivity across frequencies. This architecture guarantees exponential forgetting of the recurrent initialization and admits a finite-horizon, frequency-weighted bound on context perturbations, which extends to residual torque prediction under a Lipschitz continuity assumption on the context-conditioned dynamics model. We evaluate a low-pass realization within model predictive control on simulated and real hardware manipulation tasks, together with frequency-sensitivity analysis on real-robot data. Compared with an LSTM encoder using the same input filter, BoBDeLaN exhibits reduced high-frequency context sensitivity, lower residual torque prediction error, and smoother motion, with less degradation under injected disturbances.
Figures & tables
Figure 1: Context-Aware Deep Lagrangian Model Predictive Control (CaDeLaC) using the BoBDeLaN (dashed red box). The Pre-Filter first normalizes and the filters the history h . The MPC computes the torques τ from the state xk using the nominal model and the contextual DeLaN. The REN can be replaced by any contracting encoder with bounded incremental gain.
Figure 2: Frequency sensitivity of the context z and predicted torque τ for the LSTM and REN, estimated over 12 real-robot runs. A sinusoidal disturbance with amplitude Asj and frequencies ranging from 0.2 to 24 Hz is applied to one joint torque input j at a time, where A=0.5 and sj denotes the standard deviation of the input. Dashed curves indicate REN reference curves ( γ=0.5 ), computed using the gain ∣W(f)∣ of the second-order low-pass input filter at fb=2Hz . The references on τ additionally account for Lz , the maximum norm of ∂τ/∂z over the real-robot data.
Figure 3: CaDeLaC results for REN and LSTM encoders in simulation for unseen payloads in the training distribution, 600 runs per model. Metrics are averaged over the 7 joints.
Figure 4: CaDeLaC results for REN and LSTM encoders in simulation for payloads out of training distribution ( mL∼U(5,8) kg), 600 runs per model. Metrics are averaged over the 7 joints.
Figure 5: Pick-and-place task results for CaDeLaC with REN and LSTM encoders, evaluated with and without residual torque disturbance injection.
Appendix figures & tables5 assets
Supplementary material from the paper’s appendix.
Appendix
Parameter
Values
Activation
Tanh
Batch Size
2048
Learning Rate
5×10−4
Epochs
1500
nh
15
ϵ
0.1
Appendix
Table 1: Training hyperparameters.
Parameter
Values
ny
12
Control frequency
50 Hz
Discretization time
20 ms
Q
diag(100×14,50×13,103×14,500×13)
R
diag(0.2×14,1.0×13)
τmax , −τmin
[87, 87, 87, 87, 12, 12, 12]
Appendix
Table 2: Controller parameters.
Model
eˉq
eˉq˙
eˉτ
RMS (q¨)
RMS (τ˙)
RMS (z˙n)
[rad]
[rad/s]
[Nm]
[rad/s 2 ]
[Nm/s]
[1/s]
Median
Nominal
0.066
0.091
2.92
1.33
10.8
–
LSTM
0.018
0.101
1.86
3.55
69.2
4.67
REN
0.029
0.090
1.55
2.41
47.3
4.77
Mean
Appendix
Table 3: Median, mean and interquartile range of the tracking metrics over 600 simulated runs (30 environments with in-distribution payloads, 20 trajectories each). The metrics of each run are averaged over the seven joints.
Model
eˉq
eˉq˙
eˉτ
RMS (q¨)
RMS (τ˙)
RMS (z˙n)
[rad]
[rad/s]
[Nm]
[rad/s 2 ]
[Nm/s]
[1/s]
Median
Nominal
0.200
0.261
10.06
1.67
20.7
–
LSTM
0.065
0.237
4.83
3.98
94.0
5.08
REN
0.066
0.255
3.96
3.08
67.8
4.23
Mean
Appendix
Table 4: Median, mean and interquartile range of the tracking metrics over 600 simulated runs (30 environments with out-of-distribution payloads, 20 trajectories each). The metrics of each run are averaged over the seven joints.
Model
eˉq
eˉq˙
eˉτ
RMS (q¨)
RMS (τ˙)
RMS (z˙n)
[rad]
[rad/s]
[Nm]
[rad/s 2 ]
[Nm/s]
[1/s]
Nominal
0.033
0.048
2.97
0.82
12.0
–
LSTM
0.015
0.071
1.85
2.35
43.4
2.83
REN
0.026
0.042
0.97
1.13
17.4
1.44
LSTM with Disturbance
0.018
0.317
4.00
13.78
113.3
8.31
REN with Disturbance
0.025
0.055
1.31
1.70
33.6
2.26
Appendix
Table 5: Real-robot pick-and-place results averaged over the seven joints.
Task-agnostic assistive exoskeleton control based on human intention offers greater flexibility than conventional approaches that rely on predefined tasks or motion patterns. Human joint torque estimation enables task-agnostic assistance by characterizing user actions. Physics-consistent methods such as Deep Lagrangian Networks (DeLaN) have been applied to estimate the human torques in multi-user settings, but existing approaches cannot adapt to a specific user without retraining, and do not account for intermittent contacts during locomotion. We propose ExoLaN, a Context-Aware DeLaN for human-exoskeleton interaction that learns the full coupled system dynamics while adapting to changes in interaction context. ExoLaN combines temporal context with partial contact-force measurements from force-sensitive insoles to infer latent dynamics embeddings and estimate generalized contact torques. On seven unseen users performing 21 unseen tasks, ExoLaN reduces torque estimation MSE by 7% compared to a black-box baseline. Beyond inverse dynamics, ExoLaN serves as a unified model that also enables accurate forward prediction: training with a multi-step prediction loss reduces acceleration MSE by 59% and long-horizon position and velocity errors by 60% and 93%, respectively, compared with a single-step loss. Moreover, the learned latent context captures task information without explicit task labels, making it a promising signal for task-aware assistive control.
Lucas Schulze, Maximilian Schwarz, Jona Hoppe +2
Department of Computer Science, Technical University of Darmstadt, Germany · Robotics Institute Germany (RIG) · German Research Center for AI (DFKI), Research Department: Systems AI for Robot Learning
Robotic systems deployed in uncertain and dynamically changing environments often face variations in contact conditions, aerodynamic effects, and external disturbances that challenge reliable control. To remain effective under model-based control, these systems require dynamics models that can adapt to such changes, especially when direct access to complete environmental information is limited. To enable adaptability and facilitate integration with model predictive control, we propose a context-aware dynamics model based on neural ordinary differential equations, which infers environmental factors from state-action histories using a two-phase training procedure. We validate the approach across diverse robotic platforms, including a quadrotor in simulation, as well as a Sphero BOLT robot and a Fanuc manipulator in real-world experiments. The results demonstrate that our method effectively adapts to temporally and spatially varying environmental changes across different tasks. Videos are available at https://youtu.be/PY0sNyF2rqE , and the source code is available at https://github.com/syyu410-yu/context-aware-neural-ode-control.git .
Shao-Yi Yu, Jen-Wei Wang, Maya Horii +2
Department of Mechanical Engineering, University of California, Berkeley, CA, USA · Aalto University and YaiYai Ltd, Finland
State of the art quadrupedal locomotion approaches integrate Model Predictive Control (MPC) with Reinforcement Learning (RL), enabling complex motion capabilities with planning and terrain adaptive behaviors. However, they often face compounding errors over long horizons and have limited interpretability due to the absence of physical inductive biases. We address these issues by integrating Lagrangian Neural Networks (LNNs) into an RL MPC framework, enabling physically consistent dynamics learning. At deployment, our inverse dynamics infinite horizon MPC scheme avoids costly matrix inversions, improving computational efficiency by up to 4x with minimal loss of task performance. We validate our framework through multiple ablations of the proposed LNN and its variants. We show improved sample efficiency, reduced long-horizon error, and faster real time planning compared to unstructured neural dynamics. Lastly, we also test our framework on the Unitree Go1 robot to show real world viability.
the Robert Bosch Center for Cyber-Physical Systems, Indian Institute of Science, Bengaluru. · the Robert Bosch Center for Cyber-Physical Systems and the Department of Computer Science & Automation, Indian Institute of Science, Bengaluru.