The Sensing Unit deals with the acquisition of data from a plethora of sensors. We also pre-process raw data for Odometry and Localization and transform them into compact packets for smooth data transfer. The sensors are attached to the micro-controller units which sample and quantize the data and forward it to the microprocessor for further processing and data fusion. The Automation Unit deals with receiving processed commands (decisions) from the microprocessor and forwarding it to various actuators in the car for navigational control. Control Theory is used for controlling the actuators using error minimization algorithms. Brake, throttle and steering actuation is controlled for autonomous navigation.
The Sensing Sub-Division mainly deals with the acquisition, calibration, and manipulation of data from a plethora of sensors. The various sensors that we work on are as follows:
- 9 DOF Inertial Measurement Unit(IMU)
- Encoders (Incremental / Absolute)
- Radio Detection and Ranging (RADAR)
- Light Detection and Ranging (LIDAR)
- Proximity Sensors like Ultrasonic, InfraRed, optical flow sensors, etc.
- Mono camera
- Stereo Camera
The data is acquired real-time with the help of ADCs of microcontroller units(MCU).
Sensing Sub-Division works on Signal Processing and Multisensor Data Fusion for Localization and Odometry apart from PCB designing(through-hole + SMD) and power electronics.
The Automation Sub-Division mainly deals with actuators, Embedded systems, Interfacing and Control Systems.
- ARM microcontrollers (M4F- High Performance, R5F- High Safety)
- AVR microcontrollers (8-bit Atmega Series)
- Real-time Operating System (RTOS)
- Implementation of CAN (Controller Area Network) interface between ACCU (Automation Control Unit) and SGU (Secure Gateway Unit).
- Implementation of Ethernet, USART (Universal Synchronous/Asynchronous Receiver/Transmitter) interface between ACCU, Sensors, and PC.
- DMA (Direct Memory Access) through UART.
- SSI (Synchronous Serial Interface) between microcontroller systems.
- Simulation of Control algorithms using MATLAB and LabView.
- Mathematical Modelling of Physical Systems and tuning of PID gains.
- Study and implementation of Neuro-fuzzy intelligent controllers.
The intelligence team works along with our Artificial Intelligence Division and coordinates the communication between the ROS nodes and the microcontrollers.
They also work in the field of Machine Learning for the implementation of Artificial Neural Networks for Intelligent Controllers.