← Paper Digest

RAIPAL: A Powerful and Agile Humanoid Arm for Human Tasks

Donggeon Kim, Hyungho Chris Choi, Donghyuk Choi, Jeongsoo Park, Choongin Lee, Chan Park, Jemin Hwangbo

November 2026

TL;DR A 7-DoF, 8.9 kg humanoid arm sized from published human upper-limb measurements, with a strong upper arm, a fast forearm and an elbow linkage whose gearing changes with angle. One arm holds a 20 kg payload, pitches at 20.6 m/s endpoint speed and survives 70 hammer strikes.

What they do

Humanoid arms tend to be either strong and slow or fast and weak. RAIPAL (KAIST) is a 7-DoF, 8.9 kg arm built to be both. Its force, speed and endurance targets come from published human upper-limb measurements: the cohort mean plus one standard deviation for a 1.80 m, 80 kg adult male. Each part of the arm is sized for the tasks it serves, not for uniform payload across the workspace.

How it works

  • Force-oriented upper arm. Four identical actuators use a 22:1 3K compound (Wolfrom) planetary gearbox, rated 56 Nm with 180 Nm peak while staying backdrivable.
  • Speed-oriented forearm. Three serial wrist actuators use a 7.33:1 single-stage planetary with low reflected inertia.
  • Posture-dependent elbow. A crossed four-bar linkage varies the effective reduction from 0.531 near extension to 0.875 near flexion. That gives more speed when extended (throwing) and more torque when flexed (lifting), and it moves actuator mass toward the shoulder, cutting the effective inertia there by 15%.

Results

  • Every axis exceeds its human speed target. The shoulder axes give 248-502 N endpoint force against 129-153 N targets, while wrist flexion gives 78.6 N against 79 N.
  • One arm moved 20 kg (2.25x its mass) for 60 s. Each arm moved 5 kg for 25 minutes with all actuators below their temperature limit.
  • Pitching reached 20.6 m/s endpoint speed, about 130% of the average human reference. 70 hammer strikes needed no maintenance and left no visible damage.
  • The identified thermal model keeps the arm’s sustained-load envelope above the human endurance boundary over nearly the whole load range.

Limitations

The arm ends at the wrist flange with no hand. Every demonstration replays hand-designed joint trajectories under fixed-gain joint impedance control. The arm is not yet part of a full humanoid.

Videos

Full video (YouTube)

Figures

From Figure 1: One arm manipulating a 20 kg payload, verified for 60 s.
From Figure 1: One arm manipulating a 20 kg payload, verified for 60 s.
From Figure 1: The same arm in a high-speed upper-limb motion, verified at 20.6 m/s endpoint speed.
From Figure 1: The same arm in a high-speed upper-limb motion, verified at 20.6 m/s endpoint speed.
From Figure 2: Exploded view of the upper-arm actuator with its 3K compound planetary drive.
From Figure 2: Exploded view of the upper-arm actuator with its 3K compound planetary drive.
Abstract

This paper presents humanoid manipulator hardware designed specifically for human-level performance throughout its broad domain of tasks, from heavy payloads to highly dynamic motions. Unlike conventional designs targeting uniform workspace capability, the proposed approach derives force, speed, and endurance targets from published measurements of task-relevant human upper-limb performance. These targets guide the actuator, mechanism, and thermal design toward torque transparency and sustained operation. The platform integrates force-oriented upper-arm components, speed-oriented forearm components, and a posture-dependent elbow mechanism that shifts its torque-speed behavior. These components are driven by planetary actuators and compact actuator-embedded electronics. Experimental results validate broad task capability, demonstrating a peak payload of 20 kg (2.25x its own mass), an endpoint speed of 20.6 m/s, and continued operation under repeated hammering impacts. Thermal analysis further confirms that the robot's sustained operation capability exceeds the published human load--duration boundary across nearly the entire load spectrum. These findings show that hardware targets derived from human-task measurements narrow the human-to-humanoid capability gap. A supplementary video is available at https://youtu.be/xVUWFICD5G0

Summaries are written automatically from each paper's text. Figures and videos come from the paper and its project page; all credit goes to the authors.