Toward Gripper-Integrated Active Electrosense for Pre-Contact Sensing in Underwater Soft Grippers

Technical Summary: Toward Gripper-Integrated Active Electrosense for Pre-Contact Sensing in Underwater Soft Grippers

Problem Statement
Underwater manipulation is frequently hindered by degraded visibility caused by turbidity, glare, and gripper self-occlusion, which compromises the reliability of vision-based perception during the critical approach and grasping phases. While soft grippers offer compliant interaction suitable for these environments, they typically lack an onboard pre-contact cue to guide approach and closure when optical sensors fail. Existing tactile sensors provide rich post-contact information but cannot support pre-contact decision-making. Although active electrosense has been demonstrated for underwater object detection and localization, prior work largely treats electrodes as body-mounted arrays or dedicated experimental rigs rather than sensors integrated directly into a gripper for manipulation tasks.

Methodology
The authors propose and evaluate a gripper-integrated active electrosense system designed to provide a local, object-dependent pre-contact cue. The approach involves:

1. System Architecture: A drive–sense measurement architecture where a signal generator applies a square-wave excitation to designated electrodes to create an electric field in conductive water, while remaining electrodes record sensing voltages. Data is acquired via an NI USB-6210 DAQ.
2. Simulation: Electrostatic finite-element simulations (COMSOL Multiphysics) were conducted on an octopus-inspired soft gripper model. The study tested whether a conductive target moving within the grasp region produces spatially structured, position-dependent responses across a distributed electrode array, rather than a uniform offset.
3. Experimental Setup: Physical experiments were conducted in a water tank using a reconfigurable ribbed skeleton prototype that mimics the gripper's grasp geometry but allows for flexible electrode placement. A suspended conductive metal sphere served as the target.
4. Parameter Sweep: The authors performed a coarse parameter sweep over excitation amplitude ($V_{pp} \in \{5, 10, 20\}$ V) and frequency ($f \in \{1 \text{ mHz}, 1 \text{ Hz}, 50 \text{ Hz}, 100 \text{ Hz}, 1 \text{ kHz}\}$).
5. Metric: Detectability was quantified using a separation score $J(V_{pp}, f)$, defined as the Euclidean distance between the multi-channel voltage vector in empty water ($v_{empty}$) and the vector with the object present ($v_{obj}$).

Key Results

• Simulation: The simulations confirmed that a conductive target induces distinct, non-uniform, multi-peak voltage patterns across the sensor array. The relative amplitudes and ordering of these peaks change with the object's position, encoding coarse positional information within the grasp volume.
• Experimental Feasibility: Tank experiments demonstrated that object presence manifests as a spatial signature across sensors rather than a uniform voltage shift.
• Excitation Dependence: The study revealed a strong dependence on excitation settings. Both the aggregate separation score and the per-sensor signatures varied significantly with changes in $V_{pp}$ and frequency.
  • Representative results showed detectable changes at $V_{pp} = 5$ V ($f = 1$ mHz), $10$V ($f = 100$ Hz), and $20$V ($f = 1$ kHz).
  • The contrast between the baseline (empty water) and the object-present condition was not constant, indicating that excitation regime, grounding, and electrode placement are primary determinants of detectability.

Significance and Claims
The paper positions these findings as preliminary feasibility evidence rather than a claim of calibrated, robust performance. The authors explicitly state that the goal is not to re-establish that electrosense works in water, but to examine whether a gripper-integrated layout can provide a complementary pre-contact cue for soft manipulation.

The significance of this work lies in:

1. Integration: Demonstrating a drive–sense acquisition pipeline specifically for a gripper-integrated electrode configuration.
2. Observability: Providing evidence that multi-electrode readouts can encode object position and presence prior to contact, offering a solution to the visibility limitations of underwater manipulation.
3. Design Guidance: Highlighting that excitation parameters ($V_{pp}$, $f$), grounding, and electrode placement are critical design variables that must be systematically characterized to achieve reliable pre-contact sensing.

The authors conclude that these results motivate systematic follow-up studies, including controlled distance sweeps, characterization of water conductivity effects, advanced signal processing (e.g., lock-in amplification), and eventual closed-loop integration for approach correction and closure timing.