ROS-based autonomous driving platform for an SUV, developed by students at Tecnologico de Monterrey, Guadalajara Campus.
Mechatronics students at Tec work to create autonomous mobility — Tec Review / Conecta Tec.
The project is developed by Mechatronics students at Tecnologico de Monterrey, Guadalajara Campus, within the Center for the Car of the Future, on a 2002 Jeep Grand Cherokee (donated by NXP). The team built its own electronic steering system (gearbox), electronic throttle and brake control, and computer vision for lane following. All modifications are minimally invasive, which allows the vehicle to be returned to its original state. The goal is to build a smart-mobility lab that integrates mechanics, electronics, programming, and artificial intelligence.
iDrive is a modular autonomous driving system that integrates multiple sensors (ZED stereo camera, LiDAR, NXP microcontroller) with perception, planning, and control algorithms. The vehicle is capable of:
- Lane following through computer-vision line detection
- Real-time object detection with GPU-accelerated YOLOv3
- Distance measurement with a LiDAR Lite v3 sensor
- Autonomous control of steering, throttle, and braking via the I2C interface
ZED Stereo Camera
|
v
[vga_zed_wrapper]
|
+---+---+
| |
v v
[YOLO] [Lane Detector]
| |
+---+---+
|
v
[Navigation Control]
|
v
[NXP Communication]
|
v
Vehicle Actuators
System layout across three layers: perception (ROS + NVIDIA Jetson TX2 + ZED stereo camera), electronic control (NXP S32K148EVB microcontroller and motor driver), and vehicle (steering wheel, brake, and throttle actuated through OBD II).
| Component | Technology |
|---|---|
| Middleware | ROS (Robot Operating System) |
| Build System | catkin |
| Perception | YOLOv3 (Darknet), OpenCV |
| GPU | CUDA + cuDNN |
| Camera | ZED SDK |
| Languages | C++ (core modules), Python (lane detection) |
| Hardware | NXP S32K148, LiDAR Lite v3 |
Steering is actuated by a gearbox coupled to the steering wheel and driven by a motor with an encoder. The design is minimally invasive: it mounts on the original steering column without permanently modifying it.
Position tracking of the steering motor against a step reference, comparing the real system to the simulation. The controller tracks the reference with low steady-state error across the full operating range.
iDrive/
jeep_master_node/ # Master node - launch configurations
jeep_msgs/ # Custom ROS messages
lane_detector/ # Lane detection subsystem (Python)
ros_yolov3/ # YOLOv3 object detection (C++)
navigation_control/ # Decision and control logic (C++)
nxp_communication/ # Interface with the NXP microcontroller (C++)
lidarlite_node/ # LiDAR Lite v3 sensor (C++)
vga_zed_wrapper/ # ZED camera wrapper
testing/ # Test scripts and utilities
docs/ # Detailed documentation
- Ubuntu 16.04+ with ROS Kinetic (or newer)
- NVIDIA GPU with CUDA 9.1+ and cuDNN
- ZED SDK 2.x
- OpenCV 2.4.13 - 3.4.0
# Clone the repository into your catkin workspace
cd ~/catkin_ws/src
git clone <repository-url> iDrive
# Build
cd ~/catkin_ws
catkin_make
# Source the environment
source devel/setup.bash# Full mode: lane detection + navigation + vehicle control
roslaunch jeep_master_node jeep_master_node_lane.launch
# YOLO mode: object detection + navigation + control
roslaunch jeep_master_node jeep_master_node_yolo.launch
# NXP mode: controller communication only
roslaunch jeep_master_node jeep_master_node_nxp.launch| Document | Description |
|---|---|
| Architecture | Data flow, diagrams, and inter-node communication |
| Modules | Detailed description of each ROS package |
| Installation | Complete installation and dependency guide |
| Hardware | Sensor and I2C device configuration |
| ROS Topics | Full reference of published and subscribed topics |
Developed by students at Tecnologico de Monterrey, Guadalajara Campus.
Academic project - Tecnologico de Monterrey.