|What is a parking assistance system?|
|Lately, amidst continuous advancement in vehicles, many vehicles carry
parking assistance system. This system allows the automatic operation of
steering control while parking, based on information from camera image
and ultrasound sensor. Although drivers need to pay attention to the surrounding
condition, they can easily park their car into a garage or do parallel
parking by operating the accelerator, brake, selection lever based on the
voice guidance. In addition, the system automatically detects obstacles
and stops the vehicle if necessary by using highly precise ultrasound sensor.
Currently, the parking assistance system is being used by many auto manufacturers such as Prius and Lexus from TOYOTA. Volkswagen uses Park Assist, BMW uses Auto Parking System, and Honda uses Smart Parking Assist as parking assistance system.
|Linkage between vehicle trajectory and parking CAD|
|FORUM8 proposes vehicle trajectory, parking lot design, and auto parking
system as an integrated solution.
Vehicle Trajectory Drawing System, one of our products in UC-1 series of products, can calculate and draw a vehicle trajectory based on a drawing theory such as the one from "Design of vehicle trajectory of the semi and full-trailer turning at right angle (JASO Z 006-92), Society of Automotive Engineers of Japan, Inc." so that you can do driving simulation on an assumed path or existing path and the vehicle trajectory can be created effectively (Figure 1, 2). It is possible to design the parking lot while checking if there would be any problem to park into each space by linking the created drawing with our UC-1 series product, Parking Lot Drawing System (Figure 3, 4). This drawing of parking lot can be imported into UC-win/Road to create the parking lot in VR space. Moreover, the drawing can be printed out to use as a site where RoboCar(R) can run on for the simulation.
|Simulation by RoboCar(R)|
|FORUM8 is using RoboCar(R) for simulation in robot engineering field. RoboCar(R)
is the scaled car model and it can perform autonomous running and complex
movement by using the mounted device and software. RoboCar(R) is managed
by the main board based on Linux and can be operated using wireless LAN
UC-win/Road for RoboCar(R) is the cooperative system between 3D real-time VR and RoboCar(R). This system allows you to verify the movement of RoboCar(R) in real space while simulating in 3D VR environment. In addition, RoboCar(R) can be driven by using UC-win/Road drive simulator.
|AURELO (AUgmented Reality LOcalization System)|
|UC-win/Road needs to know the correct positional information to perform
the simulation of RoboCar(R) correctly. RoboCar(R) estimates the current
position by using the odometry. With the odometry, the number of rotations
of wheels and motor is counted in encoder, and the position is calculated
by the difference in the number of rotations. However, with this method,
many errors including wheel slip or variation in wheel width occur, consequently
creating margin of error according to the distance traveled.
To solve this problem, it is necessary to acquire the actual position information at fixed interval such as how GPS acquires.
AURELO (Augmented Reality LOcalization System) is the system for tracking the information of the absolute position and the direction. The information of 3D coordinate (x, y, z) and direction (ƒ¿, ƒÀ, ƒÁ) of the marker called "AR Maker" is extracted from the live video. The position and direction of RoboCar(R) can be measured in real-time by using two markers with different patterns at the same time. If the distance from camera to the marker is less than 1.5m, the measurement error becomes less then 7mm for the position, and less than 2 degree for the direction. The longer the distance from the camera and marker becomes, the less accurate the measurement will be. However, the coordinate can be measured across a wide area (Figure 5).
|Automatic parking system with RoboCar(R)|
|Currently, FORUM8 is developing the automatic parking system where you
can automatically park the RoboCar(R) into the specified parking space
or do parallel parking, as a function of RoboCar(R) (Figure 5). As the
RoboCar(R) synchronizes with its model in UC-win/Road, the scene of automatic
parking can be checked in VR (Figure 6). UC-win/Road easily enables the
recreation of the parking lot using VR, and the view point can be changed
so that the overview of automatic parking can be checked from various view
points (Figure 7).
RoboCar(R) automatic parking system uses AI (Artificial Intelligence). AI is the technology which makes the computer perform the intellectual process that humans perform. The system searches and decides a path in this case (Figure 8).
The current position of RoboCar(R) will be the position and direction of AURELO. In addition, the target parking space is determined based on the parking map of automatic parking system. If the route shifts a little, AI will adjust the route immediately by checking the correct positional information of RoboCar(R) constantly. AI thinks and acts like a human being by not only selecting an effective route from multiple paths but also by avoiding obstacles and searching for a new path.
You can do driving simulation in UC-win/Road by using a steering wheel. The force feedback function was developed as a new function in our latest development. Force feedback is the function which adds the resistance to the steering, and the sense of resistance increases according to the amount of rotation. The automatic parking system re-creates the scene of steering wheel turning automatically that is apparent in the actual automatic parking assistance system. Furthermore, the automatic parking can be started or stopped by pushing the button on the steering wheel. It allows the acceleration and braking operation as well.
In this automatic parking system, RoboCar(R) is controlled from outside. In the next stage of our development, the control system of RoboCar(R) will be mounted inside. RoboCar(R) will be able to acquire the information from AURELO, search for and decide the best path, search for available parking space based on the information of sensor such as stereo camera by itself. In other words, we are developing for it to be able to perform the automatic parking as close as possible to the reality.
|(Up&Coming '11 Midsummer Issue)|