Team SemTechno by TribeLab
Team members
Pedro Rodrigues, Francisco Oliveira
Team captain’s name
Pedro Rodrigues
Instructor(s)
Dr. José Sarmento, INESC TEC, TribeLab
Participating in the FRE since
2026
Description of the team and robot
We are a team of electrical engineering and mechatronics students with strong expertise in electronics, software development, and mechanical engineering. With our robot Floribot 4.0, featuring all‒wheel drive and a passive articulated joint, we improve hardware proven in previous events to develop robust and adaptable solutions for autonomous field robotics.
Our robot consists of a modified Traxxas E‒Maxx RTR wheelbase, adapted for autonomous field navigation. It integrates a custom sensor module mounted on its chassis, which includes two 2D LiDAR sensors and two cameras for environmental perception.
Additionally, the wheelbase has been mechanically modified to support a double Ackermann steering configuration, improving maneuverability and precision in structured agricultural environments.
Robot specifications
W x L x H (cm):
43 x 55 x 35
Weight (kg):
7
Commercial or prototype:
Prototype with modified commercial wheelbase
Total no. of wheels / no. driven wheels:
4/4
Drivetrain concept/ max. speed (m/s):
Double Ackermann 4-wheel drive/
approx. 1m/s
Turning radius (cm):
80
Battery type / capacity (Ah):
NiMH / 6000 mAh
Lipo / 4500 mAh
Total motor power (W):
300
No. of sensors internal/
external:
Sensor type:
2 2D LiDAR; 2 rgb cameras; 4 wheel encoders (hall sensors)
Controller system software description:
The robot is controlled via ROS2, with each sensor having its respective node which reads the input data and publishes the values as a ROS2 topic. That way it is possible for the different modules to communicate between them if required during the tasks.
There’s also a main node (supervisor), which is responsible to aggregate the different data from the remaining nodes and make the controller decisions for the navigation of the robot.
Controller system hardware description:
- Motor (steering): Servo motors
- Motor (wheel drive): DC motors
- Computer: Raspberry Pi Compute Module 5
- Microcontroller: RP2040 based board
- Camera: Raspberry Pi AI Cameras
- LiDAR: Hokuyo UST-10LX and UST-20LX
- Signals: Colored LEDs
Short strategy description for navigation and applications:
Navigation: The navigation will be based in the two LiDARs, one on the front and another on the back of the robot, to detect the lines. We will have a node that will understand the semantic of the field to decide if the robot is inside one line, outside the line and the current line number. The robot will have a localisation node that will be able to localise the current position of the robot in relation to the field. The robot has a custom controller for the double ackermann setup also allowing us to control the curvature radius and the linear velocity.
Perception: We trained two YOLO11n for the task 2 and 3. One of the models is responsible to detect the structures representing the plants with diseases (Task 2) and the other model detects the 3 classes of insects (Task 3). On both tasks, the perception is conducted on both sides at the same time, since we have one camera on each side covering the plant rows. When a detection is achieved on Task 2, an additional line is added to a CSV file stating the class detected, the row number of the corn field and the distance from the start. Also, there are 6 LEDs on the top of the robot, a set of red, green and yellow LEDs on each side. During Task 2, one of the LEDs will blink on the respective detection side to indicate a diseased plant. On Task 3, on each side, the respective LED will blink according to the class of insect detected on that side.
These are the commercial team sponsors & partners:
tba