View source for Projects:2020s1-1314 Advanced Driving Assistance System

ADAS or Advanced Driver Assistance System is designed to assist a driver in the process of driving and making the experience safe, comfortable and simpler. There are several variants of ADAS available on the market; some are built-in traits and others are available as add-on features.

[[File:ADAS System Architecture.png|thumb|System Architecture]]

While the ADAS system includes various components such as surround view, lane keeping, speed limit detection, predictive suspension, collision detection; the primary focus of this project includes developing a predictive suspension system that will integrate with the current Fully Active Suspension System at Supashock with the objective to add other functionalities in the near future.

The project will be divided into multiple phases where Phase 1 will be the primary goal and Phase 2 will be an extension to the system.

== Introduction ==

=== Phase 1 (Predictive Suspension System): ===

Phase 1 includes developing a predictive suspension system to integrate with the fully active suspension system developed at Supashock. The aim of this phase is to create a control system that uses active and passive sensors to understand the
upcoming road conditions (potholes, speedbumps and the like) integrating the vehicle information(speed, steering, orientation) to predict the optimal actuator position with respect to the vehicle dynamics to reduce sprung mass (vertical acceleration).

=== Phase 2 (Lane Keeping and collision detection): ===

Phase 2 includes adding extensions to Phase 1 to implement level 2 autonomy that assists in controlling speed and steering. The system will help with stop-and-go traffic to maintain the distance between the vehicle and the vehicle in front and also provide steering assistance to centre the vehicle within the lane. The ADAS project is estimated to start in February 2020 with the expectation of two students working to deliver a working prototype(for Phase 1 and Phase 2 if time allows) as an end result of the honours project. The activities that will be undertaken by the students include creating a simulation model of the test vehicle(SupaMule), use simulation software to simulate sensor data and test the designed control system for a predictive suspension system. Once the results are proven in simulation, the setup will be moved to the test vehicle for on-road testing inside the testing facility.

=== Project Team: ===

==== University Students: ====
Hongzhou Chen (a1709460)

Vinura Devinuwara (a1723881)

==== Academic Supervisor: ====
Prof Cheng-Chew Lim (School of EEE)

==== Industrial Supervisor: ====
Rajkunwar Kukreja (Supashock)

== Propasal and Literature Review ==
=== Aim ===
ADAS is designed to assist a driver, create better experience and improve:
* Vehicle performance
* Passenger safty
* Passenger comfort
For our project, we are focusing on developing a predictive suspension system, integrate with the current active suspension system and potentially adding other functionalities in the future.
=== Background and Previous Study ===
Active suspension controls the vertical movement of the wheels using actuators.
General ADAS systems already exist on the market, they include various components and functionalities such as surround view and lane-keeping...
However, the predictive suspension is something very new and popular, which plays an important role in ADAS. Current there are not many available applications but many companies are starting looking into it:
2020 Audi S8 
=== System Overview ===
* Obtain inputs from Lidar and camera (Hongzhou)
* Detectection, Classification and Analysis (Hongzhou)
* Integrating the vehicle information (Vinura)
* Predict the optimal actuator position (Vinura)

== Project Timeline ==
=== ROS Communication ===
On-going task
=== Lidar (Semester 1) ===
* Purchases and Integration (Mid year) √
* Point Could Handling (Week 1 - 8) √
* Simulation  (Week 1 - 8) √
* Ground Segmentation  (Week 1 - 8) √
* Speed Bump Detection  (Week 1 - 16)
* Pothole Detection  (Week 9 - 16)
* Obtaining Amplitude and Distance (Week 1 - 16)
* Surface Elevation and Other Detection  (Week 9 - 16)
* Testing and Validation  (Semester 2)
=== Image sensors (Semester 2) ===
* Machine Learning (Semester 1)
* Object Detection (Semester 2)
* Lidar and Camera Data Fusion (Semester 2)
* Testing and Validation (Semester 2)
=== Active Suspension and control (Semester 1) ===
* Understanding the passive suspension (1-2 weeks) √
* Variables of suspension that affects comfort and control (1-2 weeks) √
* Control methodologies – (Model-based and Model-free)(4-8 weeks) √
* Analysis of each control strategy(4-8 weeks) √
* Optimization based on-road profiles (1-2 weeks)
* Testing and Validation of results(1-2 weeks)

=== Predictive Suspension Control & Integration (Semester 2) ===
* Neural Network Approach (1 - 6 weeks)
* ROS communication to bind the two systems (4 – 6 weeks)
* Test and validation (4 - 6 weeks)
*

== Project Progress ==

[[File:Carla server.png|frameless|center]]
[[File:Carla classification.png|frameless|center]]
[[File:Carla rgb.png|frameless|center]]
[[File:Carla depth.png|frameless|center]]
[[File:Depth.jpg|frameless|center]]
[[File:RGB.jpg|800px|frameless|center]]

== Project Result ==
N/A
== Reference ==
N/A