Student Projects
Contact Information
University | SungKyunKwan University(SKKU) | Team Name | SAVE | |
Team Members | ChanHo Park, DoUi Hong, HwaYoung Lee, DongYeon Yu | Faculty Advisers | SungHo Hwang | |
Email Address | Contacts | 010-5044-4791 | ||
Project Information
Title: Scenario Reproduction Soft Target Development for Simulation of Unmanned Autonomous Vehicle
Description: We developed the self-driving soft target which can drive autonomously with pre-planned GPS path and scenario. The soft target can drive itself by recognizing around circumstances with Camera, GPS, and Gyro sensor and handling these acquired data with NI-MyRIO and LabVIEW.
Products:
| HW / SW | Model Number / Module |
Controller | NI myRIO | myRIO-1900 |
Development System | NI LabVIEW myRIO 2013 | LabVIEW Real-Time Module LabVIEW Vision Development Module |
Gyroscope | NI myRIO Mechatronics Kit (PmodGYRO) | L3G4200D |
Vehicle Chassis | Unmanned solution ERP (including Servo motors, MCU, LCD pannel) | ERP-42 AM-SLCD420 Serial LCD |
WebCam | Microsoft LifeCam | HD1393 |
GPS | AscenKorea GPS | AKBU3 |
The Challenge: Soft targets are used to actual pre-planned driving tests such as ADAS(Advanced Driver Assistance System) and autonomous vehicle driving test. But performing the pre-planned test with manpower might be inefficient and it may not be able to exactly perform the pre-planned scenario. To overcome these problems, we developed the self-driving soft target which can drive autonomously with pre-planned GPS path and scenario. The soft target can drive itself by recognizing around circumstances with Camera, GPS, and Gyro sensor and handling these acquired data with NI-MyRIO and LabVIEW.
The Solution: Overall soft target system is shown as the diagram below. The system is composed of three parts; those are recognition, determination, and control. First, at the recognition stage, recognize current status of vehicle from the collected data from the sensors. Second, at decision stage, generate vehicle driving path and compensate sensor error signal based on the data obtained in the recognition stage. Third, at the control stage, control movement of the vehicle such as steering angle and speed based on the decision stage. Data communications with each other at each stage are displayed as the diagram.
Figure 1. Soft Target System Configuration
1. Recognition Stage
First, the soft target must accurately recognize status of the vehicle and the surroundings in order to be self-driving. The sensors used in recognition are shown as Figure 2(a). GPS sensor recognizes current position and heading of vehicle and input desired driving path. Gyroscope is used to realize inclination status of vehicle and slope of driving road. Web camera detect lane around the vehicle and help GPS to recognize accurate current location. Data communications between hardware and controller are explained as figure 2(b).
Figure 2. (a)Hardware configuration (b)Platform-Controller signal
2. Decision Stage
Second, at decision stage, software generate vehicle driving path based on the input desired driving path and obtained around environment data from recognition stage.
Figure 3. Sensor Data Compensation Loop
Figure 3 shows compensation loop of vehicle heading and GPS location which is an example of decision making LabVIEW code. At part A, compensate GPS heading data using slope of detected lane. And then, synchronize generated driving path with input desired path at part B.
The Front panel configuration of system is shown as Figure 4 which can be divided into tab control container, image/map display, and variable list.
Figure 4. Front panel configuration
At tab control container, user can select auto/manual driving mode and determine initial vehicle speed. Image/map display show global, local map which describe information of region of interest(ROI), input reference driving path, detected lane, and location of vehicle and current camera input data at topmost display. Real-time obtained/calculated variable values show up at variable list area.
3. Vehicle Control Stage 
Figure 5. Vehicle Control Loop
Lastly, software command condition of vehicle such as speed, steering angle based on reference path and calibrated current location of vehicle. Vehicle control LabVIEW code configuration is shown as Figure 5. Key factors of the vehicle control are the distance from the target path, the corrected current position, the lane recognition, and the slope of the road.
Results
Experiment is performed at SKKU Natural Science Campus with pre-planned driving path as shown in Figure 6.
Figure 6. Experiment performed image
