A research team has designed a nested control architecture combining robust backstepping with neural estimation to improve trajectory tracking for skid-steering mobile robots in logistics applications. In a study published in Scientific Reports, researchers Khalid A. Alattas, Sahbi Boubaker, and Francisco Rossomando introduced the control strategy to address operational challenges that emerge when mobile robots perform transport duties. By uniting an outer loop for kinematic posture control with an inner loop executing robust dynamic control, the architecture provides a comprehensive framework to enhance trajectory tracking accuracy, speed, and robustness during automated material handling.
The operational requirements of logistics demand that skid-steering mobile robots execute precise and swift trajectory tracking to guarantee safe and efficient operation across dynamic environments. However, accomplishing reliable tracking is complicated by the physical dynamics inherent to this class of vehicles. Skid-steering robots exhibit highly nonlinear dynamics, nonholonomic constraints, and payload variations as cargo is transported. Furthermore, their operational performance is sensitive to external disturbances and wheel slip, factors that disrupt movement and limit the ability of classical control schemes to ensure sufficient performance.
A primary technical bottleneck identified in the investigation is the underactuated nature of skid-steering robotic platforms. Mechanically, the robot possesses three state variables but operates with only two physical control inputs. This discrepancy creates a critical control limitation by preventing direct regulation of the vehicle’s lateral error. Classical controllers often fail to deliver sufficient performance under these conditions because the robot cannot command lateral corrective forces independently. Addressing this underactuated configuration without sacrificing trajectory precision represents the central structural challenge tackled by the authors.
To overcome this structural limitation, the proposed nested system divides control responsibilities between two interconnected loops. The outer loop focuses on kinematic posture control, managing the geometric alignment of the mobile robot. Beneath this outer structure, the inner loop carries out robust tracking control by integrating a backstepping technique with an adaptive neural estimator. To compensate for the lack of direct lateral actuation, the authors formulated a virtual control law within the backstepping design. This virtual law indirectly stabilizes the lateral error through kinematic coupling with the robot’s angular velocity.
Beyond structural underactuation, skid-steering robots encounter unmodeled physical uncertainties caused by differing ground surfaces and terrain interactions. In logistics environments, vehicles experience variable lateral friction, longitudinal resistive forces, and opposition torques that alter unpredictably depending on the terrain. To compensate for these fluctuating dynamics, the authors integrated a neural estimator based on radial basis function neural networks into the inner loop. This estimator approximates unmodeled resistive forces and friction parameters, allowing the backstepping controller to adapt dynamically to changing floor conditions.
The authors established the theoretical stability of the nested control scheme through mathematical analysis using Lyapunov theory. This proof verified the practical stability of the integrated architecture, rigorously demonstrating that longitudinal, lateral, and angular tracking errors remain Uniformly Ultimately Bounded under the proposed control law. This boundedness guarantees that all spatial deviations and orientation errors stay strictly confined within acceptable limits during operation. By providing formal mathematical guarantees alongside adaptive estimation, the formulation ensures predictable performance even amidst unmodeled system dynamics.
The effectiveness and robustness of the adaptive neural robust controller were evaluated through numerical simulation across several operating scenarios. The simulation experiments incorporated simulated external disturbances and parameter uncertainties to test the system under adverse conditions. In these comparative evaluations, the proposed architecture achieved improved tracking accuracy and reduced error residuals compared to conventional feedback linearization and pure backstepping controllers. These numerical results confirm that combining radial basis function neural estimation with backstepping control provides superior tracking fidelity for skid-steering mobile robots in logistics.
Subject of Research: Trajectory tracking control for skid-steering mobile robots in logistics applications
Article Title: Trajectory tracking of SSMR using backstepping control with neural estimation for logistics applications
Article References: Alattas, K. A., Boubaker, S., & Rossomando, F. (2026). Trajectory tracking of SSMR using backstepping control with neural estimation for logistics applications. Scientific Reports. https://doi.org/10.1038/s41598-026-78031-1
Image Credits: AI Generated
DOI: 10.1038/s41598-026-78031-1
Keywords: skid-steering mobile robots, SSMR, trajectory tracking, backstepping control, radial basis function neural networks, RBF-NN, neural estimator, underactuated systems, kinematic posture control, Lyapunov theory, uniformly ultimately bounded, logistics automation, lateral friction, longitudinal resistive force, opposition torque, wheel slip, feedback linearization, virtual control law
Cite Scienmag News
Cassandra Pierce. (November 16, 2026). Neural Backstepping Controller Improves Trajectory Tracking for Skid Steering Logistics Robots. Scienmag. https://scienmag.com/neural-backstepping-controller-improves-trajectory-tracking-for-skid-steering-logistics-robots/
Cassandra Pierce. "Neural Backstepping Controller Improves Trajectory Tracking for Skid Steering Logistics Robots." Scienmag, 16 November 2026, https://scienmag.com/neural-backstepping-controller-improves-trajectory-tracking-for-skid-steering-logistics-robots/. Accessed 20 November 2026.
Cassandra Pierce. "Neural Backstepping Controller Improves Trajectory Tracking for Skid Steering Logistics Robots." Scienmag. November 16, 2026. https://scienmag.com/neural-backstepping-controller-improves-trajectory-tracking-for-skid-steering-logistics-robots/








