About development test of ADAS (Advanced Driving Support System)

This page explains the basic information on the test necessary for developing the ADAS function.
ADAS is a collective term for driving support functions for safely operating cars.
Each function implemented as ADAS, such as ACC (Adaptive Cruise Control) and FCW (Forward Collision Warning)
Based on information about recognition perception, we encourage various attention to the driver, sometimes operating the car instead of the driver.
Therefore, at the development stage, testing that assumes various situations is necessary.
Therefore, it is necessary to have a test environment that is easy to adjust to the requirements tailored to the target function and can adapt to technological innovation and legal system change expected in the future.

1. About the environment surrounding the development of ADAS

For specific functions of ADAS such as ACC and FCW, on-board ECU makes judgment based on sensor information, presents information to driver, and controls car. When developing these functions, it is necessary to assume various situations that can occur in actual traffic environments.
In addition, development requires a period of several years, but since technological innovation such as sensors will be expected to proceed during that period,
We will be able to apply the latest technology as much as possible.
Furthermore, there is a possibility that the condition for operating ADAS may change depending on the change of laws and regulations.
In this way ADAS development environment involves many changing elements, so tolerance to them is necessary.

2. Types of ADAS test

There are three types of test environments for ADAS.
 · Run a car on a program in virtual space.
· Use a miniature model or the like.
. Incorporate the necessary equipment for the actual car, and make a test on the test course and public roads.

Early stages of development are carried out on simulation, but at the final stage we will verify using actual cars on test courses and public roads.

3. About model-based development

In recent years, automotive electronics has been advanced, and 50 to 100 in-vehicle ECUs are installed in one automobile.
Model-based development is common for the development method of in-vehicle ECU. Model-based development is a method of improving the quality of algorithms and software by creating a model as a specification and repeating various tests within the simulation of the model. With the introduction of model-based development, the development period can be drastically shortened.

The process of model-based development is divided into a design phase and a verification phase. In the first half of the design phase, we will perform system design and software design. In the second half, after coding based on design, go through the steps of software verification, conformance and evaluation at the verification phase.
Model-based development benefits allow you to logically check the design content for each process, minimizing the risk of defects in the final process.

In the case of an in-vehicle ECU that realizes various functions of ADAS, we will actually verify with the hardware, which is external sensor such as camera, rider and radar, at the verification stage of model-based development.

4. Specific elements in ADAS development

ADAS has unique functions such as ACC, FCW, AEBS, TSR, LKAS and so on.
There are various sensor technologies that recognize the external world used for each function, judgment technology such as Image Recognition technology, and automobile operation technology such as steering operation, and tests are required for each function.

The challenges for these tests are:
· There is no unified hardware architecture.
· The amount and type of ECUs affected by the system is large.
· With the development of sensing technology, data complexity and capacity increase are progressing.
and so on.

5. Image of model creation

For ADAS development model elements, roads are needed first. In addition to the shape elements such as lanes, intersections, and convergence, a road model close to actual driving with signals and traffic signs is required. Furthermore, there are objects that exist on the street with moving elements such as passenger cars, trucks, motorcycles, bicycles and pedestrians, and immovable elements such as buildings and parked vehicles as external elements.
There are models of sensors such as in-vehicle radar, in-vehicle camera, rider, etc. installed in the car that detects them next.

Model creation is a difficult task.
For example, the model of the actual road has a complicated shape to be seen in a three-dimensional intersection or an interchange of a highway.
In fact, instead of creating map data from the beginning, we will generate a road model from existing map data such as "ADAS RP".

6. HIL required for ADAS test

In recent years, the need for HIL (Hardware-in-the-loop) testing is increasing as a test method for development validation of ADAS function.
HIL is a method to more efficiently test the embedded system. When testing an embedded system, it is not realistic to conduct testing after the system is completed from the viewpoint of safety and cost. By simulating the task beforehand with HIL, you can thoroughly test ECUs, sensors, control devices, etc. in the virtual environment before actual driving test.

Due to the evolution of sensor fusion technology in recent years, ADAS system is becoming more complicated.
By utilizing HIL, even if the system to be tested becomes complicated, it is possible to introduce highly reliable systems with excellent cost-effectiveness at an early stage.






8-1. 公道での走行データサービス支援

RoboTest (Robotest) is available as a Vehicle data measurement service centered on public roads, such as prototype sensor evaluation of ADAS, measurement of driving posture of the driver, data measurement of arterial roads and expressways. It is possible to measure data necessary for evaluation of prototype and research and development.
In addition, I will correspond tests such as driving tests on test courses and market research run abroad.
If you are interested, I would like to see the service introduction page below.
Travel data measurement service
RoboTest®​ ​


This service utilizes RoboCar® MiniVan, an Autonomous Driving technology development platform based on a commercially available hybrid minivan that has been marketing since 2015 as an experimental vehicle, and uses laser scanners, monocular cameras, stereo cameras, stereo cameras RoboVision® to sense the surroundings. We built a system such as GPS, inertial sensor that measures position and behavior, and in-vehicle computer IZAC® equipped with artificial intelligence that integrates acquired sensor data and performs Autonomous Driving.

For details, please see the link below.
Product and service​ ​
development support
ADAS / Autonomous Driving development support service

8-3.RoboCar MiniVan 公道実験支援パッケージ

This package utilizes Autonomous driving technology and supports customer's demonstration experiment.
For experimental vehicles, we used RoboCar® MiniVan, an autonomous driving technology development platform based on a commercial car hybrid minivan.
A laser scanner that monitors the surroundings, a monocular camera, a stereo camera RoboVision®, a GPS and inertial sensor that measures the position and behavior of the vehicle, and we have built and installed an in-vehicle computer IZAC® equipped with artificial intelligence that integrates acquired sensor data and performs autonomous operation.
For details, please see the link below.
RoboCar® MiniVan
Public road Experiment Support Package
Running on public roads using Autonomous Driving vehicles
Support demonstration and experiment


As a new service of the actual test run service "RoboTest" we are doing public road data acquisition support service with a large bus. In addition to the contents of the traveling data acquisition service so far, this service collects and evaluates data on buses and bus drivers that are difficult to arrange, based on the actual data collected by the bus and knowledge gained through acquiring traveling data of passenger cars We will support data acquisition.
For details, please see the link below.
Bus public road data
Acquisition support service
Public road travel data by large bus
Acquisition support service