Quickstart: Before You Begin

Read this first. The Paxini Gen3 is a high-density pressure and force sensor — not a motion sensor. This page explains what it measures, which grippers it mounts on, connection options (USB-C or wireless), and what you will have running when you finish setup.

What the Paxini Gen3 Measures

The Paxini Gen3 is a high-density tactile pressure and force sensor system designed to mount on robotic fingertips, finger pads, or palms. This is an important distinction from motion sensors or force-torque wrist sensors:

  • What it measures: Distributed contact pressure (spatial map), total contact force magnitude, contact area, and contact centroid location — at up to 500 Hz.
  • What it does NOT measure: Joint angles, arm pose, or velocity. It does not know where the gripper is in space — only what forces are at the contact surface.

This means the Paxini Gen3 is a sensing modality that augments your robot's gripper, not a standalone controller. Its primary value is measuring the quality of contact during grasping — capturing slip events, contact area, and pressure distribution that cameras and proprioceptive sensors cannot observe.

The most important use case: Mount Gen3 sensors on your gripper fingertips, then record manipulation demonstrations. The tactile signal tells your policy whether a grasp is stable — information that vision alone cannot reliably provide, especially for deformable or transparent objects.

What Does It Mount On?

Paxini Gen3 uses an adhesive + screw-mount adapter system that fits any flat or curved fingertip surface. No custom machining required for standard grippers.

Orca Hand

Fingertip variant mounts directly on each Orca finger pad. Full-hand tactile sensing with 5 sensors synchronized over a single USB hub. Ideal for dexterous manipulation research.

Recommended

Allegro Hand

Fingertip and finger-pad variants attach to the flat facets of each Allegro finger. Adapter brackets available from Paxini. Works with both v3 and v4 hardware.

Verified

Robotiq 2F-85 / 2F-140

Palm variant mounts on each finger face. Single-sensor setup covers the primary contact surface. Common in industrial pick-and-place and quality-inspection use cases.

Verified

Custom / 3D-Printed Grippers

The adhesive mounting system works on any clean, flat surface. Sensor thickness is 2.1 mm — thin enough to embed in most custom fingertip designs without clearance issues.

Compatible

OpenArm + Gripper

Pair with any OpenArm-compatible gripper. The sensor data streams over USB alongside OpenArm joint data for synchronized recordings via the platform SDK.

Supported

Standalone Sensing

Use without a robot arm to study human manipulation, measure contact pressure during product testing, or collect grasp reference data for imitation learning.

Supported

USB-C Wired and Wireless Options

The Paxini Gen3 supports two connection modes:

USB-C (Recommended)

Each sensor module connects via USB-C to the host PC or a USB hub mounted on the robot wrist. The sensor appears as a standard USB HID device — no driver installation required on any OS. Multiple sensors (e.g., all five fingers) connect through a USB 3.0 hub and appear as separate devices.

Round-trip latency from contact to Python callback is approximately 2–4 ms over USB-C at 500 Hz sampling. This is sufficient for detecting slip transients (<10 ms duration) in real time.

Wireless (BLE 5.2)

The Gen3 Wireless variant replaces the USB-C port with a BLE 5.2 radio and a 150 mAh rechargeable battery (8h runtime). Wireless sampling rate is limited to 200 Hz. Latency is approximately 10–15 ms, which is acceptable for data collection but not for real-time control loops.

For robot integration, use USB-C. Wireless is useful for human hand studies and bench testing. The cable management overhead on a robot arm is small — route the cable along the arm and secure with cable clips or spiral wrap.

How Long Does Setup Take?

Total time from unboxing to live tactile heatmap: approximately 1.5 hours for a single-sensor setup. Multi-finger setups add 20–30 minutes per additional sensor.

Sensor mounting
~20m
Attach Gen3 to gripper fingertip using adapter
Software install
~10m
pip install paxini-sdk, verify USB detection
First data stream
~5m
Run paxini.stream(), see pressure heatmap update
Robot integration
~45m
Sync sensor stream with arm joint data and camera

Total standalone setup: ~35 minutes. Integration with a robot arm for synchronized recording adds another 45 minutes for wiring, cable management, and timestamp verification.

Hardware Checklist

  • Paxini Gen3 Sensor Module (fingertip, finger-pad, or palm variant) Specify your gripper when ordering — Paxini ships the appropriate mounting adapter in the box. Contact SVRC for pricing →
  • USB-C cable (0.5 m, included) or USB hub for multi-sensor setups Each sensor uses one USB-C port. For a 5-finger setup, use a powered USB 3.0 hub mounted at the robot wrist to minimize cable bundle size.
  • Mounting adapter for your target gripper Included in the box for Orca Hand, Allegro, and Robotiq 2F series. For custom grippers, use the universal adhesive mount (included) or request a custom bracket.
  • Host PC — any OS (macOS, Linux, Windows) Python 3.10+. No custom driver required — sensor enumerates as a standard USB HID device. Linux recommended for robot arm integration (ROS2).
  • Robot arm or gripper (for full integration) Any compatible arm: OpenArm, Franka, UR series, custom. The sensor works standalone for bench testing but requires an arm for the full data collection pipeline.

What You Can Do After Setup

Stream live pressure distribution heatmap at up to 500 Hz from Python
Detect grasp contact onset, slip, and stable-hold events in real time
Record synchronized tactile + arm joint state + camera data during demonstrations
Publish multi-modal LeRobot datasets with tactile channels for contact-rich policy training
Assess grasp quality automatically — flag demos with slip events for exclusion or re-weighting
Train tactile-aware ACT or Diffusion Policy models that outperform vision-only baselines on deformable objects

Ready? Start the Setup Guide.

From unboxing to live tactile heatmap in under 35 minutes.