Tendon-driven aerial continuum manipulators (TD-ACMs) combine the maneuverability of uncrewed aerial vehicles (UAVs) with the compliance of lightweight continuum robots (CRs). Existing coupled dynamic modeling approaches for TD-ACMs incur high computational costs and do not explicitly account for aerial platform underactuation. To address these limitations, this paper presents a generalized dynamic formulation of a coupled TD-ACM with an underactuated base. The proposed approach integrates a strain-parameterized Cosserat rod model with a rigid-body model of the UAV into a unified Lagrangian ordinary differential equation (ODE) framework on SE(3), thereby eliminating computationally intensive symbolic derivations. Building upon the developed model, a robust dual-camera image-based visual servoing (IBVS) scheme is introduced. The proposed controller mitigates the field-of-view (FoV) limitations of conventional IBVS, compensates for attitude-induced image motion caused by UAV lateral dynamics, and incorporates a low-level adaptive controller to address modeling uncertainties with formal stability guarantees. Extensive simulations and experimental validation on a compact custom-built prototype demonstrate the effectiveness and robustness of the proposed framework in real-world scenarios.
Figures & tables
Fig. 1 : Schematic of the TD-ACM and key coordinate frames.
Parameter
Value
Unit
Nitinol backbone
Length ( L1 )
1
m
Diameter ( d1 )
2×10−3
m
Density ( ρ1 )
6.45×103
kg/m 3
Young’s modulus ( E )
5×1010
Pa
Poisson’s ratio ( ν )
0.33
–
TABLE I : Physical and geometric properties of the TD-ACM in the strain-based simulated model.
Parameter
CL
CG
Unit
Image resolution ( W×H )
1024×1024
500×500
px
Vertical field of view ( FoVv )
60
71
deg
Horizontal field of view ( FoVh )
60
71
deg
Focal length ( fx=fy )
886.8
350.5
px
Principal point ( cx,cy )
(512,512)
(250,250)
px
Near / far clipping planes
0.5/10
0.01/20
m
TABLE II : Intrinsic parameters of the local ( CL ) and global ( CG ) cameras in the PyBullet simulator.
Fig. 2 : State error evolution with target in local camera FoV.
Fig. 3 : TD-ACM input actuation evolution with target visible to local camera.
Fig. 4 : Feature trajectories and normalized error norm evolutions with target in local camera FoV.
Fig. 5 : TD-ACM state error evolution; target initially outside local camera FoV.
Fig. 6 : TD-ACM input actuation evolution; target initially outside local camera FoV.
Fig. 7 : Feature trajectories and normalized error norm evolution; target initially outside camera FoV.
Fig. 8 : TD-ACM CAD model designed in SolidWorks.
Fig. 9 : The customized TD-ACM experimental setup and its components.
Fig. 10 : Test A: (a) Time evolution of the advanced feature error; (b) evolution of the local camera twist; (c) time derivative of the CR generalized coordinates; and (d) four snapshots of the markers during operation.
Fig. 11 : Test B: (a) Time evolution of the advanced feature error; (b) time evolution of the local camera twist; (c) time derivative of the CR generalized coordinates; and (d) four snapshots of the markers during operation.
Fig. 12 : Test C: (a) Time evolution of the advanced feature error; (b) time evolution of the local camera twist; and (c) time derivative of the CR generalized coordinates.
This paper investigates two distinct approaches to the dynamic modeling of aerial continuum manipulators (ACMs): the decoupled and the coupled formulations. Both open-loop and closed-loop behaviors of a representative ACM are analyzed. The primary objective is to determine the conditions under which the decoupled model attains accuracy comparable to the coupled model while offering reduced computational cost under identical numerical conditions. The system dynamics are first derived using the Euler--Lagrange method under the piecewise constant curvature (PCC) assumption, with explicit treatment of the near-zero curvature singularity. A decoupled model is then obtained by neglecting the coupling terms in the ACM dynamics, enabling systematic evaluation of open-loop responses under diverse actuation profiles and external wrenches. To extend the analysis to closed-loop performance, a novel dynamics-based proportional-derivative sliding mode image-based visual servoing (DPD-SM-IBVS) controller is developed for regulating image feature errors in the presence of a moving target. The controller is implemented with both coupled and decoupled models, allowing a direct comparison of their effectiveness. The open-loop simulations reveal pronounced discrepancies between the two modeling approaches, particularly under varying torque inputs and continuum arm parameters. Conversely, the closed-loop experiments demonstrate that the decoupled model achieves tracking accuracy on par with the coupled model (within subpixel error) while incurring lower computational cost.
Niloufar Amiri, Shayan Sepahvand, Iraj Mantegh +1
Department of Mechanical, Industrial, and Mechatronics Engineering, Toronto Metropolitan University, Toronto, ON M5B 2K3, Canada · Aerospace Research Center, National Research Council of Canada (NRC), Montreal, QC H3T 2B2, Canada
This paper presents an AI-enabled cascaded hybrid vision/force control framework for tendon-driven aerial continuum manipulators based on constant-strain modeling in SE(3) as a coupled system. The proposed controller is designed to enable autonomous, physical interaction with a static environment while stabilizing the image feature error. The developed strategy combines the cascaded fast fixed-time sliding mode control and a radial basis function neural network to cope with the uncertainties in the image acquired by the eye-in-hand monocular camera and the measurements from the force sensing apparatus. This ensures rapid, online learning of the vision- and force-related uncertainties without requiring offline training. Furthermore, the features are extracted via a state-of-the-art graph neural network architecture employed by a visual servoing framework using line features, rather than relying on heuristic geometric line extractors, to concurrently contribute to tracking the desired normal interaction force during contact and regulating the image feature error. A comparative study benchmarks the proposed controller against established rigid-arm aerial manipulation methods, evaluating robustness across diverse scenarios and feature extraction strategies. The simulation and experimental results showcase the effectiveness of the proposed methodology under various initial conditions and demonstrate robust performance in executing manipulation tasks.
Department of Mechanical, Industrial, and Mechatronics Engineering, Toronto Metropolitan University, Toronto, Canada · Department of Mechanical, Industrial, and Aerospace Engineering (MIAE), Concordia University, Montreal, Quebec H3G 1M8, Canada
Continuum manipulators mounted on multi-rotor UAVs enable compliant aerial manipulation, but payloads and propeller downwash amplify out-of-plane bending and twisting that degrade end-effector pose accuracy. To address this problem, we present a universal-joint-based continuum manipulator designed to improve resistance to out-of-plane deformation during aerial manipulation. The proposed design uses a tubular backbone with spring-reinforced universal joints and an integrated conduit for internal routing and fluid delivery. We evaluate the design in still air and under peak propeller downwash across varying payloads, and benchmark it against a prior Nitinol-backbone CM. Bench tests show improved resistance to out-of-plane deformation across all conditions. Under peak downwash, the proposed design reduces mean error by 2.5-4x in yaw, 2-45x in y-axis, and up to 5x in roll compared to the NiTi-backbone design. We further analyze hover stability through in-flight coupled-disturbance tests over varying payloads and actuation speeds, and demonstrate the system in water sampling, spot spraying, and object transport.