Source-linked AI summary
Towards Physical Underwater Robotic Assistance for Scuba Diver Movement in Confined Spaces
Demetrious T. Kutzke, Junaed Sattar
TL;DR
Divers in confined or hazardous underwater settings need to maintain safe distance from surrounding objects, including coral. RADMCS is a wearable robot that estimates distance and uses thruster force-feedback to guide a diver’s movement; underwater studies found that 10% of maximum thrust produced a perceptible effect and that the robot could correctly indicate navigation when distance estimates were stable.
Problem
Divers in confined underwater settings must maintain safe distance from surrounding walls and objects, and contact with coral can devastate coral health.
Source: §1 · pp. 1–2Method
Results
10% of maximum thrust produced a perceptible effect, and with stable distance estimates the robot correctly indicated navigational information to the human.
Source: §6 · p. 18Takeaways & Limitations
RADMCS shows promise as a foundational platform for robotic-assisted underwater navigation, with further testing needed for more complex behaviors.
Source: §7 · p. 18In an ocean test, forward thrust shifted the device and created a painful head pinch point, moving the regulator and hoses too far forward for extended use.
Source: §5 · p. 18
Abstract
from arXiv · showhide
Scuba divers are taught to control their depth to avoid rapid ascents and descents, which could result in serious injuries such as gas embolisms and barotrauma. However, many underwater tasks necessitate lateral control, maintaining distance between subsea structures such as coral reefs, submerged drilling instrumentation, or unexploded ordnance. In this work, we discuss a first-of-its-kind wearable robotic solution providing thruster-actuated directional guidance to a diver, as distinct from prior propulsive-assistance exoskeletons. We introduce ``Robotic Assisted Diver Movement in Confined Spaces'' (RADMCS), a wearable robot that assists divers in maintaining a fixed distance from subsea structures by leveraging perception techniques in monocular depth estimation and force-feedback from submersible thrusters to provide haptic feedback. Its small and compact form factor creates a foundational platform that could be expanded to include more sophisticated control and navigation behaviors. We present results from Institutional Review Board (IRB) in-water studies with eight human scuba diver participants on threshold sensitivity tests in both a closed-water swimming facility and ocean environments; distance-maintaining experiments in a closed-water facility; and form, fit, and function testing in the ocean. We demonstrate that relatively low thrust values (10 percent of maximum) allow robotic direction of a human's movement using the physical sensation of the robot's guidance.
1 Introduction
Divers in confined or hazardous underwater environments need to maintain safe distance from surrounding objects. RADMCS addresses this with a tank-mounted wearable robot that uses thruster feedback to guide movement and is evaluated in underwater studies.
1 Introduction
- RADMCS estimates distance to surrounding structures and uses thruster-driven force feedback to perturb the diver’s movement toward maintaining a preset standoff.The robot straps onto the scuba tank and does not interfere with natural movement beyond the feedback perturbation.Source: §1 · pp. 2–3
- The work introduces a wearable underwater robotic system for movement assistance, intended to address the need for safe standoff distances around walls and objects.The motivation includes coral contact, which can devastate coral health.Source: §1 · pp. 1–3
- The authors test threshold sensitivity for haptic feedback with human participants in both ocean and closed-water environments.They also explore simulated wall-following in closed water and evaluate form, fit, and function near coral reefs in open-water field experiments.Source: §1 · p. 3
2 Literature Review
The literature review places RADMCS at the intersection of wearable haptic guidance and underwater human-robot interaction. It contrasts its tank-mounted directional assistance with prior wearable guidance devices and underwater exoskeletons.
Wearable Guidance and Assistive Technology
- Wearable guidance commonly uses vibrotactile cues from body-mounted actuators to encode direction, while other systems use phone vibration patterns or handheld auditory and vibrational feedback.Visual navigation displays can be inconvenient when users also need environmental awareness.Source: §2 · p. 3
- Garment-level electrical components and exoskeleton power systems may be problematic underwater if a suit floods, while additional cables can introduce pinch points or entanglement hazards [16].The review notes that underwater deployment of wearable SMAs or SMPs for exoskeleton or rehabilitative robotics had not succeeded to the authors’ knowledge.Source: §2 · pp. 3–5
Exoskeletons Underwater
- Underwater exoskeleton research has focused on rigid metal structures or soft cable-based and flexible designs.Prior systems include swimming-assistance devices and suits that replace or tightly couple to diving equipment.Source: §2 · p. 4
- RADMCS mounts on the scuba cylinder rather than directly coupling to the diver’s body, and its tank integration avoids the cable-and-pulley issues described for soft exoskeletons.The authors report that RADMCS costs USD 1K assembled, compared with roughly USD 100K for the atmospheric suits discussed.Source: §2 · pp. 4–5
3 System Design
RADMCS mounts to a diver’s scuba tank and combines an external attachment assembly with an internal electronics and sensing assembly. Its thrusters provide directional haptic cues, with interchangeable configurations that orient thrust across or along the robot’s body.
3 System Design
- The robot’s external assembly straps to the scuba tank, while its internal assembly houses electronics and sensing technology.The external shell is 300 mm long, 102 mm in diameter, and manufacturer depth-rated to 100 m; the tank interface can be adapted to other cylinder dimensions.Source: §3 · p. 5
- Interchangeable thruster configurations direct force perpendicular to the robot’s body in the transverse setup or along its length in the longitudinal setup.The two configurations direct thrust along the robot’s x- and y-axes, respectively.Source: §3 · p. 5, Figure 3 · p. 6
- A single 14.8 V, 6500 mAh battery powers the computational module, thrusters, microcontroller, and cameras.The stated worst-case runtime is 20 minutes at maximum thrust, while intermittent haptic actuation often exceeds four hours.Source: §3 · p. 6
- Thruster actuation provides haptic feedback by physically moving the diver in the intended direction.The system’s two thrusters are controlled by PWM signals, with neutral at 1500 µs and full forward or reverse thrust at 1900 µs or 1100 µs.Source: §3 · pp. 6–7
- Opposing forward and reverse thrust across the two thrusters signals a left or right yaw, and the directional mapping is the same for both thruster configurations.Left yaw uses reverse on T1 and forward on T2; right yaw reverses that pairing.Source: §3 · p. 7, Table 1 · p. 7
4 Visual Control Methodology
The visual control method estimates distance to a target from camera images, then converts the setpoint error into bounded thruster commands. Filtering, a deadzone, and saturation address unstable measurements and constrain assistance.
4 Visual Control Methodology
- The controller computes target distance from an image, filters successive estimates, and converts the difference from the setpoint into thruster PWM commands.The design is intended to work with any distance-estimation method, though the experiment uses classical methods.Source: §4 · pp. 7–8
- Underwater distortion, lighting changes, and autofocus can destabilize PnP distance estimates, so an EMA filter and maximum-likely-distance threshold reject large jumps.The filter uses α = 0.5; observed estimates sometimes exceeded +4 m.Source: §4 · pp. 7–8
- The controller leaves thrusters neutral inside a deadzone, then clips distance error to the sensing range before computing bounded PWM outputs.The camera and PnP method had a useful range of 1.5 m; TMIN and TMAX set the command bounds.Source: §4 · pp. 8–9
- After clipping δ to Drange, the algorithm computes T1 and T2 from δ; if δ < 0 it returns [T2, T1] for a CCW turn, otherwise [T1, T2] for a CW turn.Source: §4 · p. 8
5 Human Study Evaluations
Eight divers took part in IRB-approved in-water studies of threshold sensitivity and distance-maintaining guidance. Thresholds were low, while distance assistance depended on reliable estimates and the ocean fit test revealed a painful equipment-displacement issue.
Human Study Evaluations
- Eight divers with varying experience completed staged IRB-approved tests of threshold sensitivity and single-sided wall-following in closed-water and ocean settings.The study used a method-of-limits procedure for threshold tests, then simulated wall-following; the pool setup used blackout masks and safety-diver monitoring.Source: §5 · pp. 9, 11, 13, Table 2 · p. 10
Closed-water Threshold Sensitivity
- Across four participants and three closed-water configurations, averaged ascending and descending detection thresholds were approximately 10% of maximum thrust, with no substantial torque-arm difference.The study did not collect sensitivity data for the 325 mm longitudinal configuration; Figure 5 shows threshold results for the 225 mm longitudinal and transverse configurations.Source: §5 · pp. 10–11, Figure 5 · p. 12
Closed-water Single-Sided Distance-maintaining
- Distance-maintaining assistance worked with stable estimates, but distance loss around 12 s left filtered control stale or in a sustained high-left-thrust state.In P2’s run, cues directed the diver toward the wall when too far away and away from it when too close.Source: §5 · pp. 13–14
Open-water Threshold Sensitivity Testing
- In ocean sensitivity tests, two participants per configuration perceived motion at 10% of maximum thrust despite moderate waves and currents; the cause of this result remained inconclusive.The tests used 325 mm longitudinal and transverse configurations, with wave heights of 0.25–0.5 m.Source: §5 · pp. 14, 16
Open-water Form, Fit, and Function Qualitative Experiments
- In a one-participant ocean form-and-fit test, the device shifted forward and pinched the diver’s head, though the diver reported swimming assistance with small position adjustments.The test used the 325 mm longitudinal configuration at 80% of maximum thrust; the shift also moved the regulator and hoses too far forward for extended use.Source: §5 · pp. 16, 18
6 Discussion & Limitations
Underwater testing found that 10% of maximum thrust was perceptible and that thrusters could indicate navigational information when distance estimates were stable. The study included eight participants, and the authors call for broader testing and more complex control strategies.
6 Discussion & Limitations
- 10% of maximum thrust was required to produce a perceptible effect, and stable distance estimates allowed thruster control to indicate navigational information.The tests included threshold-sensitivity testing in all four thruster configurations and simulated wall-following.Source: §6 · p. 18
- Eight participants took part, and the logistical complexity of underwater human studies limited the study's participant count.The authors say more comprehensive testing is warranted, including evaluation of deep learning-based depth estimation.Source: §6 · p. 18
- The linear control algorithm with saturation could be extended to model-based control that accounts for both distance and human feedback.The authors identify this as an opportunity for more complex control dynamics.Source: §6 · p. 18
7 Conclusion
RADMCS is a wearable robot intended to help divers maintain safe standoff distances underwater. The authors describe it as foundational work, while noting that further testing is needed and identifying broader navigation tasks and control strategies for future work.
7 Conclusion
- RADMCS is designed to assist divers in maintaining safe standoff distances in confined or hazardous underwater environments.The authors describe the proposed system as promising foundational work in robotic-assisted underwater navigation.Source: §7 · p. 18
- Further open-water threshold-sensitivity testing with more participants is warranted, alongside broader navigation tasks and evaluation of threshold generality.The planned work also includes refining control strategies for simulated confined-space traversal.Source: §7 · p. 18
- The authors hope the work will attract broader community investment in safety-critical wearable robots underwater.Source: §7 · p. 19