Electrical
Subsystem

01

Overview

The electrical subsystem of the Jellybot is a multi-tier power and communication architecture designed to bridge high-current actuation with precision digital control. At its core, the system regulates power for a dual-Raspberry Pi configuration, coordinating the Master–Slave communication required for remote underwater surveillance.

To meet the specific demands of coral reef monitoring, the subsystem prioritises actuation stability over raw thrust. It integrates a BlueOS Flight Navigator with a 2-pump and 2-solenoid valve array, allowing for a dampened response characterised by a time constant τ^[1]. This configuration ensures that the electrical load remains optimised at 76.8 W, extending mission endurance to 3.1 hours.

02

Electrical System

Three main system configurations were evaluated before arriving at the adopted design. Each configuration differs in the number of actuator pumps and whether a flight controller is integrated, directly affecting power consumption, endurance, and control authority.

Table 1 — System Configuration Comparison

2.2 Power Profiles

Each component's voltage, current, and power draw were characterised individually before computing total system loads. All pump and valve connections run in parallel on a 12 V supply rail.

Table 2 — Individual Component Specifications Table 3 — System Power Profiles and Endurance (12 V, 20 000 mAh battery)

03

Wiring Diagrams

Wiring diagrams were produced for all three system configurations. Each configuration shares a common power regulation stage but differs in the number of pump and valve branches on the 12 V actuation rail. Click any diagram to view it full screen.

04

Tests & Results

To determine the optimal configuration for the coral detection mission, two key tests were conducted: a pump configuration test evaluating actuation performance at the target 71° tentacle bend angle, and an energy consumption test quantifying mission endurance across all configurations.

4.1 Pump Configuration — 2 vs 4 Pumps

To demonstrate that a 2-pump configuration can match the steady-state performance of a 4-pump system, both setups were evaluated at maximum actuation. A maximum tentacle bend angle of ~71° was confirmed using HX710 pressure sensing at 206 kPa^[1] — sufficient to overcome hydrostatic pressure beyond 5 m depth.

The bend angle tracking software was used to monitor deformation underwater in real time. Both configurations were driven until reaching steady-state at 71°, but with differing transient behaviour.

Table 4 — 4-pump & 2-pump data

Both systems achieved steady-state at 71°. The 4-pump system reached it in 0.7 s (τ ≈ 0.45 s), while the 2-pump system reached it in 1.4 s (τ ≈ 1.1 s, measured at 63.2% of steady-state = 45.1°).

4.2 Velocity Analysis

Five test runs were recorded for each configuration to characterise horizontal propulsion velocity. The 4-pump system achieved significantly higher peak speed, but this came at the cost of stability for close-range imaging.

Table 5 — Comparative Velocity Analysis of Jellybot Propulsion

The 2-pump system's σ = 0.14 indicates a high level of consistency across cycles. While the 4-pump configuration offers superior raw speed (10.0 cm/s peak), it proved unsuitable for high-resolution coral reef imaging — at 10.0 cm/s, surge velocity introduces motion blur and reduces the window for stable visual detection. Furthermore, the high-volume displacement of 4 pumps causes excessive vertical lift, making the robot prone to "overshooting" its target depth during actuation.

4.3 Depth Stability & Precision

A critical requirement for underwater surveillance is the ability to hover at a fixed altitude. During active depth-control tests, the following depth oscillations were observed for each configuration:

Table 6 — Depth Deviation Comparison During Active Depth Control

The 2-pump system's slower time constant (τ) effectively acts as a "dampener,"^[1] allowing finer granular control over buoyancy. This reduction in vertical jitter is essential for maintaining the constant focal length required for automated coral detection algorithms. A higher τ describes a more sluggish but smoother system which in the coral reef context, this is a feature, not a limitation.

4.4 Conclusion

Table 7 — Conclusion of Data

The adoption of the 2-pump configuration (System 3) is justified by three key engineering factors:

05

Battery & Power Endurance

To determine the optimal configuration for the coral detection mission, four electrical power profiles were evaluated using a 12 V, 20 000 mAh battery. The primary metrics were power draw (Watts) and operational endurance (Hours).

Table 6 — Comparative Power Profiles and Operational Endurance

While System 2 offers maximum theoretical runtime (3.2 hr), System 3 was adopted as the final architecture. The BlueOS Flight Navigator adds only 2.4 W (a negligible 6-minute reduction in runtime) while providing the critical computational overhead for depth-stability algorithms, real-time video feedback, and live telemetry.

The On-Board Pelican Case at 4.0-hour battery life comfortably exceeds the Jellybot's 3.1-hour endurance, ensuring the surface station remains operational throughout the full deployment, recovery, and data-offloading phases.

06

Future Work

07

List of Figures

08

References

1. Cisneros-Limón, R., & Castro, E. (2023). Soft fluidic closed-loop controller for untethered underwater gliders. arXiv. https://doi.org/10.48550/arXiv.2303.08672

2. Zhu, J., Jin, G., Li, S., & Zhang, W. (2023). Effect of low-pass filtering on passivity and performance of series elastic actuation. IEEE Robotics and Automation Letters, 8(6), 3326-3333. https://ieeexplore.ieee.org/document/10115477