RC UANL RoboCup Junior Rescue Maze About us RC-UANL is a team from Mexico, representing FIME and the SOFTWARE HARDWARE Universidad Autónoma de Nuevo León, that competes in the Rescue Maze category at TMR 2026. It integrates Mapping Algorithm Electronics robotics expertise with advanced technologies in The robot's mapping system is based on a DFS (Depth-First Search) algorithm Microcontroller responsible H-bridge module to autonomous systems to develop a robot capable of combined with an exploration strategy guided by absolute references, RGB color sensor to for receiving information control the speed and Teensy 4.1 APDS9960 detect black, blue, red, TB6612FNG navigating complex environments, detecting victims, and allowing for more consistent and predictable navigation within the maze. from the sensors and direction of the DC and silver floors. overcoming obstacles, aiming to improve its performance The environment is represented as a dynamic matrix, where each cell executing movements. motors. throught the navigation. corresponds to a tile. As the robot advances, it records relevant information IMU that provides DC motors that offer Diego Emilio Evangelista Gomez such as walls, available paths, detected victims, and special zones. This Programmable camera for absolute orientation. good torque and representation allows for the identification of explored and pending areas, computer vision. Responsible Designer OpenMV H7+ BNO055 Responsible for Micromotor N20 speed. They feature for victim detection. Responsible of mechanical design. optimizing decision-making. controlling turns and encoders to calculate Unlike traditional strategies such as the "right-hand rule," the robot prioritizes inclinations. the distance traveled. its movements based on an absolute orientation system. An initial direction is Stefany Giselle Colunga Rangel High-current step-down Distance sensor with a Binary sensor that, defined as absolute north, determined by the robot's orientation at the start of voltage converter. range of up to 4m. Programmer along with the bumper the execution. From this reference, the exploration priority order is: D24V50F5 Responsible for powering 5V VL53L0X Responsible for Micro Limit Switch Responsible of Mapping Algorithm components. sensing walls. design, allows for the Absolute North detection of obstacles. Absolute East Rogelio Andrade López Absolute South Programmable RGB Limited-rotation servo Electronics Absolute West Low-dropout (LDO) linear LED strip/ring. motor. Along with a Responsible of PCB design. At each intersection, the algorithm evaluates the available directions voltage regulator. Responsible for rack and pinion AMS1117 WS2812 SG90 Responsible for powering providing light and mechanism, it is following this order, always selecting the unexplored option with the highest 3.3V components. indicators during the responsible for Axel Uriel Labrada Hernández priority. This allows for the generation of more uniform trajectories and round. dispensing kits. Programmer facilitates the consistency of the generated map. Responsible of victim detection. When the robot reaches a point with no new available routes, the algorithm executes a backtracking process, returning to previously visited nodes until it PCB Previous Results finds new unexplored branches, this ensuring complete coverage of the The robot's electronic system is integrated through a custom PCB, designed to environment. centralize connections and optimize space usage within the chassis. The Track Strategy The team participated in the board was developed considering the efficient distribution of all components, Robocup Mexican Open 2025, including sensors, actuators, and communication modules. The design also reaching a 3rd place at a In the first run, the robot avoids the red zone due to its high complexity, considers the proper management of communication protocols such as I2C national level. prioritizing stable navigation and returning to the starting point to obtain the and UART. This allows for reduced wiring and improved internal organization. Winners of the Mexican Open exit bonus. Thanks to this integration, the PCB acts as the core of the robot, ensuring a 2026 If a lack of progress occurs after a checkpoint, this objective is discarded, and robust and organized connection that is adaptable to future improvements. We are currently preparing to represent Mexico at RoboCup the robot prioritizes exploring the entire map, including the red zone. The strategy aims to minimize the Lack of Progress (LoP) and maximize the Mechanics 2026, with the goal of score through adaptive decisions during execution. The mechanical design of the robot is aimed at ensuring stable and adaptable reaching an international podium finish. Victim Detection and Delivery navigation in complex environments. A compact structure with approximate dimensions of 16x14x14 cm was developed, allowing for better control within Platforms and Languages An OpenMV camera handles real-time victim detection using a resource-efficient, dual- the maze. Additionally, a low and centered center of gravity was prioritized, approach system: achieving greater stability and reducing the risk of tipping over when Letter Victims: A lightweight MobileNetV2 neural network classifies alphabetical characters overcoming obstacles such as ramps and speed bumps. on the maze walls. Cognitive Targets: A geometry-based Computer Vision algorithm processes targets without heavy neural networks through three strict stages: Suspension System Detection & Proximity Validation: Uses color blob detection and geometric filtering to The suspension is designed as a mechanical system that allows a certain identify perspective-distorted ellipses. To comply with the 15 cm deployment rule, analysis only triggers if the ellipse's minor axis exceeds a safe pixel threshold. degree of rotational movement in the wheels through axles and movable LAB Space Classification: Samples pixels across 5 concentric rings. By calculating the joints. When the robot traverses irregularities such as speed bumps, ramps, or squared Euclidean distance to pre-calibrated LAB centroids, it classifies colors robustly steps, the suspension absorbs these variations, allowing the wheels to adjust against varying venue lighting: Blue (2), Green (1), Yellow (0), Red (-1), or Black (-2). to the terrain. This prevents the loss of contact with the surface and maintains Health Status Evaluation: Sums the 5 ring values to determine the exact kit requirement the robot's stability. (Harmed = 2, Stable = 1, Unharmed = 0). Invalid mathematical sums are instantly rejected as false positives. Bumper More information 2. Communication & Actuation UART Protocol: The OpenMV encodes the required kit count (0-2) and transmits it to the main The robot features a front bumper system made of flexible material, controller via UART (pins 2 & 5) for fast, reliable synchronization. Inventory & Kit Management: The robot features two side towers holding 4 rescue kits each. integrated with limit switches. This mechanism allows it to detect collisions Based on the UART signal and the target's location, the main controller activates Neopixel with walls or obstacles, facilitating position correction and reducing indicators, actuates the corresponding tower's servomotor to drop the kits, and updates its accumulated errors during navigation. internal inventory counter accordingly. Wheels Movement Functions The tires were designed to maximize traction on different surfaces. They are Advancement between tiles is achieved through controlled 30 cm made of silicone and feature a pattern that allows them to adapt to obstacles movements, utilizing encoders to measure the distance traveled and ensure such as ramps and stairs, improving the robot's performance on variable precision in every movement. To maintain the correct orientation, an IMU- terrain. based correction system is employed, allowing for precise 90° and 180° turns. Furthermore, position correction routines are implemented using distance Kit Ejection System GITHUB and contact sensors, allowing the robot to align itself with the walls and reduce accumulated errors. The kit delivery system is based on a servo-driven rack and pinion mechanism, designed to be compact and precise.