MochiSwarm: A testbed for robotic blimps in realistic environments

TL;DR

MochiSwarm introduces an open-source, modular testbed of lightweight robotic blimps designed for autonomous swarm operation without external localization. The system combines a compact hardware stack (XIAO ESP32S3 MCU, differential-drive servo-rotor modules, detachable NiclaVision perception) with a software framework based on a state-machine and ESP-NOW communication, enabling onboard sensing and centralized coordination. Perception modules detect target balloons and retroreflective goal hoops using color-based detectors and shape checks, supporting visual servoing for autonomous capture and delivery. Experiments demonstrate near-linear scalability in workload with increasing swarm size and successful autonomous MAPD operation under indoor wind disturbances, validating MochiSwarm as a flexible platform for developing and evaluating multi-robot aerial systems.

Abstract

Testing aerial robots in tasks such as pickup-and-delivery and surveillance significantly benefits from high energy efficiency and scalability of the deployed robotic system. This paper presents MochiSwarm, an open-source testbed of light-weight robotic blimps, ready for multi-robot operation without external localization. We introduce the system design in hardware, software, and perception, which capitalizes on modularity, low cost, and light weight. The hardware allows for rapid modification, which enables the integration of additional sensors to enhance autonomy for different scenarios. The software framework supports different actuation models and communication between the base station and multiple blimps. The detachable perception module allows independent blimps to perform tasks that involve detection and autonomous actuation. We showcase a differential-drive module as an example, of which the autonomy is enabled by visual servoing using the perception module. A case study of pickup-and-delivery tasks with up to 12 blimps highlights the autonomy of the MochiSwarm without external infrastructures.

Figures (7)

• Figure 1: We test the MochiSwarm system in an industrial highbay. The blue blimps capture red target balloons and deliver them through light-yellow hoops. A video summary is available at https://youtu.be/75gfIlPtrBU.
• Figure 2: The hardware components of a Mochi blimp.
• Figure 3: A close-up picture of the Mochi frame. The arrows show the direction of inserting the latch pieces to secure the frame, the central openings in blue host the battery, the sensor stack, and the main PCB from top to down, the red opening resides the servo-rotor stack, and the yellow traces host elastic strings that connect to the LTA balloon.
• Figure 4: The software architecture of the Mochi firmware. The bold arrow marks the entry of the firmware, which instantiates the state machine that specifies the autonomy of the blimp, depending on the task, the blimp actuation model, and the sensors available on the hardware construct.
• Figure 5: Left: the image frame captured by the NiclaVision where The pink blocks highlight the activated grid cells that match the balloon color. Right: the color distribution in the LAB color space for balloons with different colors.