Flight through Narrow Gaps with Morphing-Wing Drones
Authors: Julius Wanner, Hoang-Vu Phan, Charbel Toumieh, Dario Floreano
Organizations: Laboratory of Intelligent Systems, École Polytechnique Fédérale de Lausanne (EPFL), Lausanne, Switzerland · Avian-Insect Robotics in Motion Laboratory, University of Nevada, Reno, USA
The size of a narrow gap traversable by a fixed-wing drone is limited by its wingspan. Inspired by birds, here, we enable the traversal of a gap of sub-wingspan width and height using a morphing-wing drone capable of temporarily sweeping in its wings mid-flight. This maneuver poses control challenges due to sudden lift loss during gap-passage at low flight speeds and the need for precisely timed wing-sweep actuation ahead of the gap. To address these challenges, we first develop an aerodynamic model for general wing-sweep morphing drone flight including low flight speeds and post-stall angles of attack. We integrate longitudinal drone dynamics into an optimal reference trajectory generation and Nonlinear Model Predictive Control framework with runtime adaptive costs and constraints. Validated on a 130 g wing-sweep-morphing drone, our method achieves an average altitude error of 5 cm during narrow-gap passage at forward speeds between 5 and 7 m/s, whilst enforcing fully swept wings near the gap across variable threshold distances. Trajectory analysis shows that the drone can compensate for lift loss during gap-passage by accelerating and pitching upwards ahead of the gap to an extent that differs between reference trajectory optimization objectives. We show that our strategy also allows for accurate gap passage on hardware whilst maintaining a constant forward flight speed reference and near-constant altitude.
Figures & tables
Fig. 1: A wing-sweep morphing drone flies through a narrow gap by temporarily sweeping in its wings.
Fig. 2: The 130 g morphing-wing drone platform leveraging wing-sweep morphing for narrow gap passage. (a) The outboard wing sweep is actuated by a servo-driven bar-linkage mechanism. The sweep angle Θw varies between -5° and 75° from the fully extended to the fully swept configuration respectively. (b) The side profile view shows the outboard flat-plate sweeping into the symmetric root airfoil.
Fig. 3: (a) Aerodynamic characterization setup using ATI Gamma load cell, a wind generator and a robot arm positioning the drone at angles of attack between -8° and 90°. We compare the derived aerodynamic model to measurements of (b) lift coefficients, (c) drag coefficients, and (d) pitch moment coefficients at a sweep angle of -5° (extended wing configuration) and 75° (fully swept wing configuration) at an inbound wind speed of 5 m/s. (e) Lift coefficient measurements and aerodynamic model predictions at wind speeds of 4 and 6 m/s at a wing sweep angle of 45° are additionally shown.
Fig. 4: Description of the morphing-winged drone narrow gap passage optimization problem: The drone approaches from level flight at a specified initial flight speed. During a chosen gap anticipation distance, the drone can vary its symmetric wing sweep, thrust and elevator inputs freely to preempt the narrow-gap passage (Phase 1). During a specified gap threshold distance, the symmetric wing sweep must remain in a fully swept actuator state and the drone must maintain an altitude corresponding to the center of the narrow gap (Phase 2). After gap passage, the drone is given a free distance to recover to a constant flight condition at the altitude corresponding to the gap center, with free choice of symmetric-wing sweep, elevator and thrust (Phase 3). In the final phase (Phase 4), the drone resumes steady flight at the initial launch altitude with a freely chosen constant velocity, attitude and wing sweep configuration.
Fig. 5: Optimization framework and control architecture, showing the generation of the reference trajectory using multi-phase optimization, the Nonlinear Model Predictive controller to generate symmetric sweep, motor thrust and elevator servo commands, and the low-level lateral attitude controller computing asymmetric wing sweep and rudder corrections to maintain lateral alignment to the gap.
Fig. 6: Comparison of optimal reference trajectories among three flight optimization objectives: minimum variation of altitude across all flight phases (Case 1), minimum variation in forward flight speed (Case 2), and minimum time during the gap passage phase (Case 3). For the exemplary case of an initial level flight speed of 6 m/s, a gap location at 4 m, and a gap threshold of 0.8 m, we show: (a) Altitude relative to the launch altitude, (b) flight speed relative to approach speed with wing sweep input overlay (100 % represents fully swept wings), and (c) drone pitch angle with throttle input overlay. Boxplots display the the following for every flight optimization objective summarized across three initial level flight speeds (5,6 and 7 m/s), gap thresholds (0.4, 0.8 and 1.2 m) and gap locations (4, 5 and 6 m): (d) Time-averaged altitude during the gap anticipation phase, (e) Gap passage flight speed relative to approach speed, (f) Time-averaged pitch angle during the gap passage phase. A Mann-Whitney U-test was performed between datasets of interest, yielding P < 0.0001 (****) between all compared selections.
Fig. 7: Narrow-gap passage flight experiments for closed-loop control with Nonlinear MPC and reference trajectories for a flight optimization objective of least altitude variation (Case 1). We visually show the gap passage phase with fully swept wings and the inward sweeping and extending process for an initial speed of 6 m/s and a gap threshold of (a) 0.4 m, (b) 0.8 m and (c) 1.2 m. Exemplified considering an experiment with an initial speed of 6 m/s and a gap threshold of 0.8 m, we show (d) the barycentric mean trajectory across three flights compared to the computed reference trajectory with the computed vertical error to the gap center and (e) the predicted mean absolute flight speed and mean pitch rate by the dynamic model over one MPC prediction time-step compared to values measured across three flight experiments. For flight experiments between initial speeds of 5, 6 and 7 m/s and gap thresholds varying between 0.4, 0.8 and 1.2 m, we compare the (f) Altitude error measured at the gap location, and the (g) Mean altitude tracking error during the recovery phase.
Fig. 8: Closed-loop control flight experiment data obtained for reference trajectory optimization objectives of minimum variation of altitude across all flight phases (Case 1), minimum variation in forward flight speed (Case 2), and minimum time during the gap passage phase (Case 3). We show the (a) altitude error measured at the gap location and (b) mean altitude tracking error during the recovery phase. A gap threshold of 0.4 m and initial speeds of 5, 6 and 7 m/s are set.
Aerial robotics studies the design, perception, control, and coordination of flying robots for navigation, inspection, transport, and physical interaction.
Aerial payload transport using multirotor unmanned aerial vehicles is challenging because payload geometry, contact interaction, grasp stability, flight control, and narrow-passage traversal are strongly coupled during pickup and transport. Object-specific grippers often cannot adapt their footprint or grasp geometry when the payload shape or passage width changes. This paper presents a vision-guided morphing quadcopter for multi-geometry payload transport through narrow passages. The proposed platform uses four hybrid arm-leg structures that function as both landing supports and grasping members. A centrally placed actuator drives a tendon-based morphing mechanism, enabling all four arms to synchronously retract or expand for object grasping, footprint reduction, and post-transport release. Onboard vision estimates the payload geometry and passage width, while endpoint force feedback is used to confirm grasp contact during payload engagement. A phase-wise mission planner, PID-based flight stabilization, and morphology-adaptive grasp controller are implemented in a MuJoCo simulation environment. The framework is evaluated using box, cylindrical, and spherical payloads, representing flat-faced, rolling-curved, and fully curved contact conditions. Across the three cases, the simulated system completes the pickup-transport-release sequence with a maximum RMS position error of 0.31 m, a final drop-zone error below 0.18 m, a compact grasp footprint of 0.09-0.21 m2, and a footprint reduction of 75.0-89.7 percent. The results demonstrate that a single-actuator morphing quadcopter can adapt its grasp footprint for the transport of payloads with different geometries while reducing its overall footprint for narrow-passage traversal.
Aashish Sahu, Shriram Hari, R. Prasanth Kumar
Department of Mechanical and Aerospace Engineering Indian Institute of Technology Hyderabad Hyderabad, India
This paper introduces a motion planning framework to plan morphology and trajectory for morphing quadrotors under extremely constrained environments. We develop a novel obstacle avoidance cost function for nonlinear model predictive control (MPC) that enables navigation through extremely narrow gaps under limited perception from a 2D LiDAR. Classical artificial potential field-based costs typically have a high cost in narrow passages, artificially blocking the navigable path. In contrast, we propose a smooth exponential obstacle cost that preserves low traversal cost within narrow gaps while maintaining strong collision avoidance behavior. The formulation avoids hard activation thresholds and introduces a cost reduction factor to reduce the cost within narrow passages. Direct use of 2D LiDAR measurements in MPC allows navigation around arbitrarily shaped obstacles. The method is embedded within an acados-based nonlinear MPC framework. Simulation and experimental results demonstrate successful traversal of narrow corridors where typical repulsive cost functions would fail. The approach provides a computationally efficient and practical solution for navigating through tight spaces while maintaining safety from the obstacles. While we are implementing the framework on the morphing quadrotors, the cost function formulation is general-purpose for any mobile robot application, and is not limited to the morphing quadrotors. The implementation code is available at Github Repo and a short video is available at Video Link.
Harsh Modi, Xiao Liang, Minghui Zheng
Department of Mechanical Engineering, Texas A&M University, College Station, TX 77843, USA · Department of Civil & Environmental Engineering, Texas A&M University, College Station, TX 77843, USA
Precise aggressive maneuvers with lightweight onboard sensors remains a key bottleneck in fully exploiting the maneuverability of drones. Such maneuvers are critical for expanding the systems' accessible area by navigating through narrow openings in the environment. Among the most relevant problems, a representative one is aggressive traversal through narrow gaps with quadrotors under SE(3) constraints, which require the quadrotors to leverage a momentary tilted attitude and the asymmetry of the airframe to navigate through gaps. In this paper, we achieve such maneuvers by developing sensorimotor policies directly mapping onboard vision and proprioception into low-level control commands. The policies are trained using reinforcement learning (RL) with end-to-end policy distillation in simulation. We mitigate the fundamental hardness of model-free RL's exploration on the restricted solution space with an initialization strategy leveraging trajectories generated by a model-based planner. Careful sim-to-real design allows the policy to control a quadrotor through narrow gaps with low clearances and high repeatability. For instance, the proposed method enables a quadrotor to navigate a rectangular gap at a 5 cm clearance, tilted at up to 90-degree orientation, without knowledge of the gap's position or orientation. Without training on dynamic gaps, the policy can reactively servo the quadrotor to traverse through a moving gap. The proposed method is also validated by training and deploying policies on challenging tracks of narrow gaps placed closely. The flexibility of the policy learning method is demonstrated by developing policies for geometrically diverse gaps, without relying on manually defined traversal poses and visual features.
Tianyue Wu, Guangtong Xu, Zihan Wang +6
Institute of Cyber-Systems and Control, College of Control Science and Engineering, Zhejiang University, Hangzhou, China. · Huzhou Institute of Zhejiang University, Huzhou, China. · Differential Robotics Technology Co., Ltd., Hangzhou, China.