Landing on a dime: the biomechanics and kinematics of lovebirds (Agapornis roseicollis) landing on a swinging perch

Birds frequently must land safely and accurately on moving branches or power lines, and seemingly accomplish such maneuvers with acrobatic precision. To examine how birds target and land successfully on moving supports, we investigated how lovebirds (Agapornis roseicollis ) approach and land on a swinging perch. Lovebirds were trained to take off from a hand –held perch and fly ~6 m to land on a servo–controlled swinging perch, driven at three sinusoidal frequencies, in a purpose–built flight corridor. Lovebird flight and landing kinematics were recorded using a motion capture system. A force–torque sensor mounted to the landing perch recorded the bird′s horizontal and vertical landing force and pitch torque. In support of our hypothesis for stable landings, lovebirds timed their landings in a majority of trials (51.3%), when the perch was approaching either extreme of its motion with its velocity nearing zero (27.5% in the same direction as the bird′s approach — SDS, and 23.8% in the opposite direction to the bird′s approach — ODs). As a result, lovebirds exhibited a robust bimodal strategy for timing their landing to the phase of the swinging perch. Less commonly, lovebirds landed when the perch was moving at high velocity either toward the bird′s approach (12.3%) or in the same direction as the bird′s approach (11.5%); with the remainder (21.9%) of trials distributed over a broad range of swing phases. Landing forces were greatest in the horizontal plane, with vertical forces more varied and of smaller magnitude across all landing conditions. This reflected the shallow (more horizontal) flight trajectory (approach angle: 31.9 ± 3.5° SEM) that the lovebirds adopted to decelerate and land. Increased landing force correlated with greater landing speed of the bird relative to the perch (R² = 0.42956, p < 0.0001). The lovebirds initiated landing with a consistent body pitch angle (81.9 ± 0.46° SEM relative to horizontal) across all landing conditions, using the horizontal perch reaction force to assist in braking when landing. Subsequent head–down body pitch rotation of the bird after landing was not well correlated and generally opposite to the initial direction and magnitude of landing pitch torque, which was generally negative due to foot rotation and ankle flexion at landing. Flexion of the birds′ hind limb joints (ankle: −29.2 ± 9.2°, knee: −13.6 ± 7.4°, and hip: −4.0 ± 3.4° at landing, combined with their horizontal approach trajectory, reduced the magnitude of landing torque by aligning the bird ′s center of mass trajectory more closely to the landing perch than if they landed from above the perch. Landing pitch torque and body pitch rotation also increased uniformly in response to increased perch swing frequency. In contrast to landing forces, landing pitch torque was more varied across landing conditions, as well as in relation to the phase of landing. In general, higher landing force was encountered when the perch was moving towards the approaching bird. Our results indicate that birds regulate their approach trajectory and velocity to time the phase of landing to a moving perch, providing insight for designing biologically–inspired unmanned aerial vehicles capable of landing on moving targets.