AI-Enabled Image-Based Hybrid Vision/Force Control of Tendon-Driven Aerial Continuum Manipulators

A team of researchers has published a paper detailing a novel AI control system for a new class of flying robots. The system is designed for 'tendon-driven aerial continuum manipulators'—essentially drones equipped with soft, flexible, tentacle-like arms. The core innovation is a hybrid controller that fuses visual data from an onboard camera with force feedback from a sensor. This allows the drone to 'see' an object and 'feel' the contact force simultaneously, enabling precise physical interactions like pushing, pulling, or inserting objects in unstructured environments.

The controller's intelligence comes from a combination of two key AI techniques. First, a 'fast fixed-time sliding mode control' algorithm provides robust stabilization against disturbances. Second, a 'radial basis function neural network' performs rapid, online learning to compensate for uncertainties in both the visual scene and the force measurements. Crucially, this learning happens in real-time during operation, eliminating the need for lengthy offline training sessions. For vision, the system uses a state-of-the-art graph neural network to extract 'line features' from the camera feed, a more reliable method than traditional geometric extractors.

In comparative benchmarks, the proposed AI framework demonstrated superior performance against established methods used for drones with rigid arms. Simulation and physical experiments showed the system could successfully execute manipulation tasks under various starting conditions, proving its robustness. This represents a significant step beyond simple aerial observation, moving drones into the realm of active, dexterous physical work in complex settings where both perception and delicate touch are required.