Continuous in-hand rotation requires coordinated contact maintenance and renewal, linking manipulation performance to both hand geometry and tactile feedback. We investigate these relationships through rotational progress, finger participation, and fingertip loading on two versions of the low-cost LEAP Hand with anthropomorphic and symmetric finger arrangements. Both hands are integrated with TacTip vision-based tactile sensors and use a shared reinforcement learning framework. Simulation evaluates full tactile, binary touch, and proprioception-only observations on 12 unseen objects; hardware experiments evaluate full tactile policies on ten objects under multiple hand orientations. In simulation, full tactile observations increase mean normalized rotation per episode over proprioception alone by 94.8% for the anthropomorphic hand and 66.9% for the symmetric hand, with longer mean episode durations and lower mean total fingertip loads. Hardware contact measurements reveal more differentiated finger participation in the anthropomorphic hand, while cumulative loading is more evenly distributed across the symmetric hand's fingers. Overall, the symmetric hand achieves a 32.0% higher mean rotation count and a 25.8% higher aggregate rotation rate, together with longer manipulation duration. Additional hardware force recordings show a 19.9% lower mean total fingertip load for this layout. These findings relate finger arrangement to contact organization and load sharing, and show that richer tactile information supports greater task progress at lower mean loading in both simulated embodiments.
Figures & tables
Figure 1: Overview of the approach. Left: the two LEAP Hand embodiments. Right: teacher–student training and deployment. The teacher is trained in simulation using deployable observations and privileged information. The student is trained under teacher supervision using the current deployable observation and a latent estimate of privileged information inferred from a finite history of deployable observations and commands. The trained student policy is deployed on the corresponding physical hand.
Figure 2: Hand orientations used in real-world evaluation for the Anthropomorphic and Symmetric LEAP Hand embodiments.
Table I: Simulation performance under different observation conditions.
Hand orientation
Rotation count (revolutions) ↑
Manipulation duration (s) ↑
Anthropomorphic LEAP Hand
Palm Up
1.20±0.20
28.38±3.03
Palm Down
1.01±0.29
25.88±6.14
Thumb Up
0.94±0.56
30.00±0.00
Thumb Down
0.88±0.22
27.40±4.92
Base Up
1.04±0.30
22.92±7.08
Table II: Real-world manipulation performance across hand orientations.
Figure 4: In-hand manipulation with the Anthropomorphic LEAP Hand under different hand orientations.
Figure 5: In-hand manipulation with the Symmetric LEAP Hand under different hand orientations.
Figure 6: Fingertip-centre workspaces for the anthropomorphic (left) and symmetric (right) hands. Orange, red, and blue denote cells reachable by exactly two, three, and four fingertips, respectively, across sampled configurations. Of 40,000 sampled points, 8,000 are displayed.
Metric
Thumb/ Finger 1
Index/ Finger 2
Middle/ Finger 3
Little/ Finger 4
Anthropomorphic LEAP Hand
Contact time (%)
91.23
81.62
93.12
75.98
Load share (%)
28.99
22.40
28.24
20.37
Transition rate (s -1 )
1.101
2.454
0.875
2.261
Mean force during contact (N)
4.13
3.57
3.94
3.48
Symmetric LEAP Hand
Table III: Hardware finger participation and fingertip loading.
Metric
Full tactile
Binary touch
Proprioception only
Anthropomorphic LEAP Hand
Mean total load (N)
10.31
12.28
13.49
Cumulative-load CV
0.138
0.135
0.085
Symmetric LEAP Hand
Mean total load (N)
11.63
11.93
13.01
Cumulative-load CV
0.050
0.024
0.029
Table IV: Fingertip loading and cumulative load sharing in simulation.
Metric
Anthropomorphic LEAP Hand
Symmetric LEAP Hand
Mean load ± SD (N)
13.01±5.09
10.42±3.37
Upper load level (N)
20.86
16.44
Cumulative-load CV
0.148
0.051
Table V: Hardware fingertip loading and cumulative load sharing.
We investigate in-hand rolling manipulation using a multifingered robot hand, where each finger is compliant and equipped with a tactile fingertip providing contact location and wrench information. We derive the equations of motion for compliant quasistatic in-hand rolling manipulation and formulate a fingertip rolling manipulation controller for multiple fingers to achieve a desired object twist within a grasp. The contact mechanics are demonstrated in simulation and the controller is tested on an experimental robot system.
Huan Weng, Yifei Chen, Kevin M. Lynch
Center for Robotics and Biosystems and the Department of Mechanical Engineering, Northwestern University, Evanston, IL 60208 USA · Northwestern Institute on Complex Systems (NICO)
Humanoid hands require tactile feedback across the whole finger, not just the fingertip, to grasp and manipulate objects properly. Vision and proprioception alone cannot reliably provide this information, particularly when the hand's own fingers occlude the camera's view of the grasp. We present a low-cost tactile array for a humanoid finger, made from Velostat and conductive tape. The fingertip carries seven contact points including a 2x3 matrix wrapped across its front, left, and right faces, and a separate contact point at the tip. The proximal and middle phalanges each carry a single front-facing contact line. We measure the sensor's hysteresis and recovery time after release, through repeated loading and press-release tests. We also test a compliant, 3D-printed contact structure with a gap and a bump, inspired by similar designs in prior work, and show it cuts recovery time by 74% compared to a flush-contact baseline. We then show the sensor can produce distinct activation patterns for different contact geometries (flat, edge, corner) at the fingertip, and that it registers contact across all three phalanges during a grip. Finally, we discuss the limits of our fabrication changes and point to software-based compensation as a promising way to more directly fix the remaining hysteresis in the future.
Neel Adwani, Muhaiminul Islam Akash, Rituja Bhattacharya +1
New Jersey Institute of Technology (NJIT), Electrical and Computer Engineering
Dexterous object rotation is a sequential contact problem: each support, release, and re-contact decision must both produce the desired object motion, and prepare the hand configuration for continued rotation. Existing reinforcement learning methods discover such movement patterns through trial and error on specific robotic hand embodiments, without explicitly accounting for how each contact transition affects the hand's ability to sustain object rotation in subsequent steps. We introduce DexMani, a framework that transfers human demonstrations as contact-conditioned manipulability evolution. This prior captures how successful human contact transitions reshape the object-rotation directions available to the hand. DexMani then learns this manipulability evolution and uses it to guide downstream reinforcement learning, enabling rotation skills to be acquired across robot embodiments with distinct kinematics and active-contact configurations. Across the Shadow Hand, Allegro Hand, and XHand, DexMani achieves the highest success rates in every evaluated setting for both seen and unseen objects. DexMani reaches an average success rate of 57.5% on LEAP Hand, outperforming other baselines and producing smoother rotatory motions. Project site: https://dexmani.github.io
Xiaoyang Chen, Shengcheng Luo, Haoran Guo +4
Shanghai Jiao Tong University · ShanghaiTech University · Beijing Institute for General Artificial Intelligence (BIGAI) +1