Projects

RaccoonBot:
An Autonomous Wire-Traversing Solar-Tracking Robot for Persistent Environmental Monitoring

Robotics · Mechatronics · Embedded Systems · Power Electronics · PCB Design · Additive Manufacturing · IoT · AI/ML · Energy Autonomy · System Architecture · System Integration

RaccoonBot is an autonomous, bio-inspired robotic platform I conceived developed for persistent environmental monitoring. Designed to traverse suspended wires while tracking the sun for energy harvesting, it integrates robotics, embedded systems, power electronics, environmental sensing, IoT connectivity, and AI/ML-based energy management into a field-ready system that was deployed at the Crystal Cove State Park in Newport Beach, California, USA, for nearly a year from 2025 to 2026.

Tasked with developing an original robotic concept for the UCI Robot Ecology Lab, I established RaccoonBot's bio-inspired design and led its engineering development from concept and system architecture through prototyping, manufacturing, integration, validation, and deployment. During its field operation, RaccoonBot collected and remotely transmitted environmental data while serving as a real-world testbed for persistent robotic autonomy.

Stage 1 – Robotic Platform Development

RaccoonBot is an autonomous, wire-traversing robotic platform designed for persistent environmental monitoring under challenging outdoor conditions. It features a self-locking mechanism, a custom-developed active solar-tracking algorithm, and an integrated solar-energy harvesting system that supports extended operation with minimal human intervention.

I conceived the original bio-inspired concept and led RaccoonBot's multidisciplinary engineering development, from system architecture and mechanical, electrical, embedded, and power-system design to prototyping, manufacturing, integration, and experimental validation. Throughout its development, I also directed and mentored team members across different project stages, coordinating their contributions toward the realization of a fully integrated robotic system.

A central engineering challenge was developing a reliable wire-traversing locomotion mechanism capable of operating on inclined cables while minimizing energy consumption. I designed a motor-driven, self-locking worm-gear transmission that enables RaccoonBot to maintain its position without continuously powering its motor. I also developed and integrated the electrical and embedded systems, including motor control, environmental monitoring instrumentation, power conversion, battery management, and solar energy harvesting. The custom solar-tracking algorithm adjusts the photovoltaic panel's orientation to maximize available charging power.

I also led RaccoonBot's hardware manufacturing and prototyping, designing and fabricating its custom PCBs, structural enclosures, mechanical attachments, and distinctive bio-inspired exterior. I extensively applied additive manufacturing, evaluating different 3D printing materials, including ABS, PETG, and others, before selecting ASA, Tough PLA, and PLA+ based on mechanical performance, environmental resistance, and manufacturability. The final design also incorporated metallic components such as stainless steel and aluminum. I additionally directed the design and fabrication of a lubricated gearbox enclosure and coordinated the acquisition of specialized 3D printing, soldering, and electrical testing equipment to support in-house production, assembly, and validation.

Through iterative mechanical, electrical, and functional testing, I validated the platform's locomotion, energy harvesting, and subsystem reliability. Following laboratory and outdoor evaluations, RaccoonBot was successfully deployed at Crystal Cove State Park in July 2025, demonstrating autonomous wire traversal, solar tracking, environmental monitoring, and embedded control under real-world conditions. The project subsequently became the flagship robotic platform and visual identity of the UCI Robot Ecology Lab, reflecting its role in the laboratory's research and outreach activities.

Publications

In this stage, the development and experimental validation of the robotic platform contributed to three publications:

• IEEE ICRA 2025 – RaccoonBot: An Autonomous Wire-Traversing Solar-Tracking Robot for Persistent Environmental Monitoring.

• IEEE Access Journal Paper – V-model Based Mechatronic and Power Electronics Integration for an Efficient Solar-Powered Wire-Traversing Robot.

• IEEE Conference Paper – Noise Mitigation in Power Sensing under EMI via RMS-Based Numerical Filtering for Embedded Systems in Mechatronic Devices.

Media Coverage

Durign this stage, RaccoonBot and the UCI Robot Ecology Lab received media coverage highlighting the platform's nature-inspired design, solar-powered autonomy, and environmental monitoring applications:

• Los Angeles Times (April 2025) – UCI’s Robot Ecology Lab Inspired by Nature.

• Energy Innovation Review (May 2025) – California's RaccoonBot: Solar-Powered Robot Revolutionizes Remote Monitoring.

Stage 2 – Edge-to-Cloud IoT Architecture

Deploying RaccoonBot for persistent environmental monitoring introduced a new engineering challenge: environmental data collected in the field needed to be reliably transmitted, stored, and accessed remotely, while the robot's essential sensing and autonomous functions remained independent of continuous internet connectivity.

To address this challenge, I designed the edge-to-cloud IoT architecture and led its multidisciplinary implementation, distributing functionality across the robot, a local gateway, and the cloud. I developed the robot-side embedded communication system, integrating environmental measurements and LoRa transmission with the main control PCB I designed. I also mentored team members in the development of a dedicated environmental sensor PCB and directed the integration of the Raspberry Pi gateway, Firebase Firestore, and web monitoring interface, coordinating communication and compatibility across the independently developed subsystems.

The resulting architecture transformed RaccoonBot into a persistent remote environmental monitoring system, connecting field measurements to a remotely accessible platform while preserving autonomous operation at the edge. Its successful integration and deployment at Crystal Cove State Park demonstrated the coordinated operation of embedded hardware, wireless communication, gateway processing, cloud storage, and remote data visualization under real-world conditions.

Beyond the technical development, I led RaccoonBot's field deployment and operational coordination at Crystal Cove State Park, representing UCI in meetings with park stakeholders and collaborating with facilities personnel to identify and prepare a suitable installation site. I designed the required deployment equipment, conducted on-site evaluations and testing, and coordinated the installation process. Following deployment, I also scheduled and performed maintenance activities to support the robot's continued operation under real-world environmental conditions.

Publication

The development and field validation of this architecture contributed to the following publication:

• IEEE Internet of Things Journal (IoT-J) — Enabling Remote Persistent Environmental Monitoring with Robotic Platforms using an Edge-to-Cloud IoT Architecture.

Media Coverage

RaccoonBot's deployment at Crystal Cove State Park and its applications in environmental conservation, research, and STEM education received coverage from local news, university communications, and conservation organizations:

• Newport Beach Independent (July 2025) – Tech Meets Tidepools with Crystal Cove's Newest Conservation Hero RaccoonBot.

• UCI Samueli School of Engineering (July 2025) – High School Students Learn About Environmental Robots at UCI–Crystal Cove Summer Institute.

• Crystal Cove Conservancy (2025) – Coverage of RaccoonBot's deployment at Crystal Cove State Park and its role in environmental monitoring and conservation.

Stage 3 – Bio-Inspired Torpor and Learning-Based Energy Autonomy

Long-term field deployment exposed one of the fundamental challenges of persistent robotics: RaccoonBot needed to remain operational despite changing weather and periods of limited solar energy. Rather than simply increasing battery capacity, I looked back to the animal that inspired the robot itself. During harsh environmental conditions, raccoons can reduce their activity and energy expenditure to conserve energy when food becomes scarce. Drawing an analogy between food availability in nature and solar energy availability in robotics, I developed a bio-inspired torpor strategy that enables RaccoonBot to conserve its limited energy resources by adapting its activity to environmental conditions.

I initially implemented this behavior through a rule-based torpor controller, integrating solar energy availability, battery conditions, and embedded power management to coordinate transitions between active and low-power operating states. Extended field operation revealed the limitations of fixed decision rules under unpredictable weather conditions, motivating me to develop a learning-based controller using an artificial neural network (ANN) trained with the Earthquake Algorithm.

I integrated the learning-based controller with RaccoonBot's behavioral state machine and embedded power-management architecture, enabling adaptive energy-management decisions based on the robot's operating conditions. Supported by nearly a year of real-world field operation, the bio-inspired energy-management framework was evaluated under changing environmental conditions. Comparative field evaluations further showed that the learning-based strategy maintained healthier energy levels, recovered from energy deficits, and exhibited greater resilience during periods of reduced solar availability than the original rule-based approach.

By combining a survival strategy inspired by nature with embedded intelligence, machine learning, and hardware-level energy management, RaccoonBot evolved beyond conventional solar-powered operation toward adaptive energy autonomy, enabling more sustainable and persistent environmental monitoring under challenging outdoor conditions.

Publication

The development and field validation of this bio-inspired energy-management framework contributed to the following manuscript:

• IEEE Transactions on Field Robotics (T-FR) (Under Second Review) – Learning-Based Energy-Aware Torpor Control for Persistent Solar-Powered Robots.

Autonomous Drone:
A Custom-Built UAV for GPS-Guided Flight and Navigation with ArduPilot

Robotics · Autonomous Systems · UAVs · Autonomous Navigation · Flight Control · Embedded Systems · Sensor Integration

Driven by my interest in autonomous systems and my passion for hands-on engineering, I took Dr. Peter Burke's Autonomous Drones course at UCI as an opportunity to explore a new robotics domain beyond my research work.

I built a fully autonomous UAV from the ground up, with the goal of developing an aerial platform capable of executing a complete GPS-guided mission without manual flight control. Starting from individual components, I assembled the mechanical platform, soldered and integrated the electronics and power system, and incorporated the onboard flight controller, IMU, GPS, ESCs, telemetry radios, and failsafe systems.

Bringing the aircraft from a collection of integrated subsystems to reliable autonomous flight required extensive calibration, debugging, control tuning, and iterative flight testing. I configured the ArduPilot firmware, calibrated the onboard sensors, and tuned the PID control loops, progressively validating the aircraft from initial flight tests to stable controlled flight and, ultimately, autonomous operation. The completed UAV successfully executed a fully autonomous GPS-guided mission, including takeoff, waypoint navigation, and landing, demonstrating the successful integration of its mechanical, electrical, sensing, control, and embedded systems.

World’s Smallest Autonomous Aquatic Robot:
A Miniaturized Robotic Platform with MHD Propulsion, Embedded Systems, and Resource-Constrained Autonomy

Robotics · Mechatronics · Embedded Systems · Power Electronics · PCB Design · Additive Manufacturing · MHD Actuation · Autonomous Systems · AI/ML · Sensor Integration

The World’s Smallest Autonomous Aquatic Robot (WSAAR) project was an ambitious research effort to miniaturize an autonomous aquatic robot capable of sensing its environment, making decisions, and navigating through water. Beyond a single robot, the long-term vision was to pave the way toward swarms of miniature autonomous aquatic robots working cooperatively, with potential applications ranging from environmental monitoring and water remediation to future microrobotic systems.

Making that vision possible meant confronting a fundamental challenge: everything becomes harder as the robot gets smaller. Space, energy, computation, sensing, propulsion, and fabrication become increasingly constrained and interconnected, requiring the entire robotic system to be reconsidered around miniaturization.

My work tackled these challenges across multiple fronts, including the development of Magnetohydrodynamic (MHD) propulsion and its driving electronics, custom PCBs and embedded systems, resin-based additive manufacturing, and resource-constrained autonomous control. These technologies were developed and validated through complementary experimental platforms, progressively advancing the capabilities needed for WSAAR and the broader vision behind it.

Stage 1 – Magnetohydrodynamic Propulsion Development

Miniaturizing the robot required a propulsion system that could generate motion without relying on conventional rotating components. We explored magnetohydrodynamic (MHD) actuation, which produces fluid motion through the interaction between electric and magnetic fields, offering a compact propulsion approach well suited to the scale targeted by WSAAR.

I worked on the design, fabrication, and experimental characterization of the MHD actuators, iterating their geometry, electrodes, and manufacturing process while evaluating their electrical and propulsion behavior. In parallel, I developed the custom power-electronic driving circuitry required to regulate the current through the actuators and reliably control their operation (as addressed in Stage 2). These iterations progressively brought the propulsion system closer to the size, power, and integration requirements of the robotic platform.

Stage 2 – Miniaturized Electronics & Robotic Platform Integration

Developing the MHD actuator was only the first half of the propulsion challenge. Making it move was one problem; developing electronics capable of driving and controlling it within the footprint of a miniature robot was another. Unlike a conventional actuator that could be paired with a standard motor driver, the MHD propulsion system required dedicated power electronics capable of precisely regulating the electrical conditions across the actuator while remaining compact and efficient enough for integration into WSAAR.

I designed custom MHD driving electronics using DC-DC converters as current regulators, creating dedicated drivers for the two actuators and achieving electrical efficiencies of up to 97% during experimental testing. These drivers then had to coexist with the robot's power regulation, sensing, and embedded computation, turning the electronics into a system-level miniaturization challenge where power conversion, control, component selection, PCB layout, and physical packaging all competed for extremely limited space.

To bring these functions together, I developed a custom PCB and embedded-system architecture integrating the MHD drivers with an ATtiny10 microcontroller, Hall-effect sensing, and power regulation. In parallel, I used resin-based additive manufacturing to iteratively prototype the robot's mechanical structure and packaging, allowing the mechanical and electronic designs to evolve together toward a miniaturized aquatic robotic platform capable of supporting sensing, control, and autonomous operation.

Stage 3 – Parallel Development of Autonomous Tracking

While the WSAAR hardware was being miniaturized, I pursued the development of its autonomous capabilities in parallel using a larger aquatic robotic platform, allowing the sensing and control strategy to be developed and experimentally validated without waiting for the miniaturized system to be completed. I designed a lightweight artificial neural network (ANN) tracking algorithm with the goal of enabling autonomous navigation under the severe computational constraints envisioned for WSAAR, using metaheuristic optimization to train the network before deploying it on the experimental platform.

The platform originated from a team I mentored in a robotics course I taught at UCI, where the students developed an aquatic robot capable of tracking a laser source. Recognizing its potential as a research platform, I brought the team into the project and mentored them through the implementation and experimental validation of the ANN-based controller on the robot they had built. This parallel approach allowed us to validate the autonomous tracking strategy while WSAAR's hardware continued to evolve, advancing both sides of the project simultaneously. The resulting work was later published as a research paper on autonomous aquatic robotic tracking:

• IEEE Conference Paper – Autonomous Aquatic Laser-Following Robot Through RGB Sensors and Optimized Artificial Neural Networks.

Self-Balancing Robot:
PSO-Optimized Control on a Custom Robotic System

Robotics · Mechatronics · Embedded Systems · FreeRTOS · PCB Design · Motor Control · PID Control · Particle Swarm Optimization (PSO) · Real-Time Control · Experimental Validation

Developed as a graduate-level control engineering project during my Ph.D. at Tecnologico de Monterrey, this two-wheeled self-balancing robot presented a challenging problem: maintaining stability in an inherently unstable system while improving its response to external disturbances. Rather than relying exclusively on conventional controller tuning, I explored how Particle Swarm Optimization (PSO) could improve the performance of a PID controller through intelligent parameter optimization.

I designed and built the complete robotic platform from the ground up, including its mechanical structure, motor-driving electronics, power supply, and a custom PCB for reliable hardware integration. I also developed the embedded software architecture using FreeRTOS on an NXP FRDM-K64F microcontroller, integrating sensor acquisition, signal processing, and real-time motor control to maintain the robot's upright position.

To improve its balancing performance, I implemented a PSO-based PID tuning methodology and designed an experimental validation protocol comparing the optimized controller against conventional Ziegler–Nichols tuning. Both approaches maintained the robot within ±5° of the upright position, but the optimized controller achieved a 97.86% improvement in mean balancing-angle error and an 86.79% reduction in angular-response variability, demonstrating substantially improved stability and disturbance rejection.

I presented the resulting work at the joint ICMEAE-2020 and ICCRE-2020 IEEE conferences in Osaka, Japan, where I received the 🏆Best Presentation Award🏆.

• IEEE Conference Paper (🏆Best Presentation Award🏆) – Self-Balancing Robot Control Optimization Using PSO.

Reconfigurable Micro-CNC Platform:
Embedded Firmware Architecure & System Integration

Embedded Systems · Firmware Development · Software Architecture · System Integration · Industrial Automation · CNC Machining · Motion Control · Motor Control · HMI Development · Mechatronics

As a Software R&D Engineer at Tecnologico de Monterrey, I contributed to the development of a reconfigurable Micro-CNC manufacturing platform for the Bi-National Laboratory on Smart Sustainable Energy Management and Technology Training. The objective was to create a modular system capable of supporting different manufacturing processes while remaining reproducible and adaptable for deployment across multiple institutions.

My primary contribution was designing and developing the layered embedded firmware architecture that coordinated the machine's operation. Starting with research into existing open-source solutions, I developed a firmware framework integrating memory management, communication protocols, G-code interpretation, and motion-control coordination. The main microcontroller communicated with dedicated motor-control boards, enabling the integration of mechanical, electrical, and software subsystems into a functional manufacturing platform.

Beyond firmware development, I led junior engineers in the integration of the HMI and peripheral systems, including the OLED display, control buttons, switches, and electrical connections. By the end of my direct involvement, the platform supported milling and turning operations, while its extensible architecture provided a foundation for additional manufacturing applications, including 3D printing and laser processing, developed through subsequent collaborative efforts.

The platform was engineered and documented for repeatable assembly, integration, and deployment, with machine kits delivered to at least 10 institutions during my involvement in the project. This approach extended the system beyond a single prototype, enabling independent teams to assemble, operate, and further develop the machines for different manufacturing applications.

ExoTec:
Wearable Lower-Limb Rehabilitation Exoskeleton with EMG-Based Control

Robotics · Mechatronics · Exoskeletons · Mechanical Design · Embedded Systems · PCB Design · Motor Control · EMG-Based Control · Fuzzy Logic Control · System Integration

As the culminating project of my B.S. in Mechatronics Engineering with a Robotics Concentration at Tecnológico de Monterrey, I led the development of ExoTec, a wearable lower-limb rehabilitation exoskeleton designed to assist knee movement and reduce the muscular effort required during rehabilitation exercises and standing motions. Starting from the initial concept, I directed the multidisciplinary development process, coordinating mechanical design and manufacturing while personally developing the system's mathematical modeling, electronics, embedded firmware, instrumentation, and control architecture.

The mechanical design was based on knee dynamics and torque calculations, supported by finite element analysis (FEA), material selection, and consultation with a rehabilitation specialist. I directed the CAD development and manufacturing of a custom structure incorporating telescopic supports, motor-driven joints, and mechanical elements designed to stabilize knee movement and support the exoskeleton's weight.

On the electronics and control side, I designed and manufactured a custom PCB, integrated DC motors, encoders, power electronics, and electromyography (EMG) sensing, and developed the embedded firmware in C using an ATmega16A microcontroller. To synchronize the two motors, I implemented a master–slave control architecture with a Mamdani fuzzy-logic controller, using encoder feedback to compensate for differences in motor motion. I also developed the signal-processing and PWM-based control routines to modulate motor assistance according to muscle activation, overcoming the microcontroller's memory and processing limitations through firmware optimization.

The resulting prototype was experimentally validated through mechanical, electrical, and functional testing, demonstrating synchronized motor operation, EMG-responsive assistance, and approximately 36 N·m of available motor torque. Additional operating modes enabled external torque adjustment and push-button activation, allowing the system to be demonstrated interactively during the university's engineering project exhibition.

Completed in approximately one semester, ExoTec became a defining milestone in my early engineering career. Its successful development attracted academic attention and contributed to opportunities that ultimately led to my first engineering position at RT Venalia and an invitation to pursue a fully funded direct Ph.D. program.