Jongsoo Kim Director Korean Auto Safety Association Korea
Establishment of autonomous driving verification process (Hyundai Motor-K-City collaborative case study)
Je Myoung Ryu Global R&D master Hyundai Motor Company Korea
Validation and verification of automotive system safety is becoming increasingly important in order to expand ADAS and commercial autonomous-driving technology. Without careful definition of required safety functions as well as preparation of proper testing methods in advance, the ultimate self-driving technology will be far from complete. We will review major differences in testing methods between conventional automotive performance and self-driving performance in view of the vehicle-human-environment interaction. Hyundai Motors R&D is focusing on three types of testing and four pillars of validation processes for accelerating advanced technology development. Among them, proving ground testing will be one of important challenges for the establishment of future mobility test environments. We will look at the status of a collaborative project with KATRI using K-City for test scenarios and test systems building. It is going to propose the establishment of a joint verification system between organizations and industries in order to respond to new NCAP and self-driving certification in the future.
Development of assessment methods for automated vehicles
Dr EunDok Lee Senior researcher KATRI Korea
According to the World Health Organization (WHO), about 1.3 million people die from road traffic accidents every year, and 90% of these accidents are caused by drivers who are poor at driving, tired or negligent. Therefore, traffic accidents could be drastically reduced if the causes of accidents by drivers were reduced. This solution is the autonomously driven car. Autonomous vehicles are one of the key keywords of the 4th Industrial Revolution and strive to gain technological advantage not only in Korea but also in other developed countries. Autonomous vehicles are a useful means of reducing social costs by reducing traffic congestion, fuel consumption and pollution emissions as well as traffic accidents. To promote early commercialization of such autonomous vehicles, Korea aims to provide evaluation technology for ensuring the safety of autonomous vehicles at Level 3 by 2020. To this end, the Korea Automobile Testing and Research Institute (KATRI) is conducting research on the development of autonomous vehicle safety evaluation technology and a testbed with domestic and overseas research institutes and universities. To improve the safety and reliability of autonomous vehicles developed by automobile manufacturers, universities and research institutes, test environments capable of repetitive and reproducible tests in a controlled environment are required. K-City is a testbed where we are trying to improve the safety of automobiles.
Status and plan of autonomous driving service development based on precision map of Daedeok Science Town
Jeongdan Choi Principal engineer ETRI Korea
Automated vehicles must satisfy basic seven functions in order to acquire a temporary driving license for autonomous vehicles defined by the Ministry of Land, Transport and Communications as ‘Regulations on safety operation requirements and test operations of automated vehicles’. The seven functions are basic driving, lane change, cut-in/cut-out, congestion follow-up and release function, forward collision prevention function, maximum speed limit function and fail-safe function. In addition to the basic functions, the pedestrian recognition or signal recognition function is tested to suit the traffic environment in which the vehicle is intended to be operated by automated vehicles. Having these basic requirements and verifying the basic functionality of the test run, it can be used to derive business models for a variety of automated driving services on public roads. Our team were authorized for temporary automated driving operation to develop automated driving services in traffic situations, including general road conditions, pedestrian crossings and signal crossings. To do this, we introduce an electronic control platform using an electric power source and a vehicle platform equipped with camera sensors and lidar for urban driving. This paper introduces a driving system that recognizes multiple objects in the traffic situation on the road and generates the major maneuver and control values based on the precision map. In addition, it introduces AI technology used for traffic-light recognition and pedestrian recognition, and location recognition technology using camera and lane markings. Lastly, it will present lessons from tests such as automated driving taxi using terminals in a general traffic situation around Daedeok Research Complex, and plans for the automated driving service in the future.
AD-STEP for Level 3, 4, 5 autonomous driving test and evaluation platform
Jungchan Oh Service development manager SpringCloud Inc Korea
AD-STEP is a government-funded project that is developing a test procedure and evaluation system based on international standards. This is targeted for Level 3, 4 and 5 autonomous driving scenarios such as highway, valet parking and urban situations.
Efficient data fusion development based on a holistic workflow
Keun Wook Jung Manager Vector Korea Korea
Accurate implementation of the driver assistance system requires accurate awareness of the vehicle perimeter. Sensor fusion technology is required to perceive the surroundings of a vehicle based on data from various sensors such as radar, smart camera and lidar. We will explain how Baselab’s tools are utilized and how they can be used to develop sensor fusion algorithms from initial prototype code to mass code generation.
13:05 - 13:35
Dr EunDok Lee Senior researcher KATRI Korea
Universal safety framework for the single approval of automated vehicles
Joong-Seok Oh Head automotive team TÜV Süd Korea
Dr Houssem Abdellatif Global head autonomous driving and ADAS TÜV Süd Germany
Min-Kyu Shim Senior Engineer HAD TÜV Süd Korea
Currently there are no general regulations or methods for certification or approval of automated vehicles. However, many use cases or initiatives are struggling to operate automated vehicles on public roads. TÜV Süd has developed a universal framework that copes with the heterogenous local regulations and provides a unified method for the assessment of automated vehicles. This framework considers local roadworthiness regulations, as well as functional safety and cybersecurity, and can therefore be applied in any region in the world. TÜV Süd has enabled ambitious projects to be successful in operating automated vehicles on public roads, without any safety issues.
Challenges of autonomous driving measurement technology validation
Davor Kovacec CEO Xylon Croatia
Hyuk Kim Xilinx Korea
Autonomous driving systems require diverse sensors that are distributed and channeled via various communication protocols. The rising sensor resolution and ever-changing high-speed communications make validation of these systems very challenging and increase requirements for reliable and safe multi-channel capturing of sensor data with synchronized and accurate time stamping, which is essential for the further use of the recorded data. The modular and scalable logiRECORDER datalogging platform, with its integrated data recording and playback capabilities, quickly adapts to differing requirements and helps overcome challenges through all development and testing stages, such as sensor validation, algorithm development and hardware-in-the-loop (HIL) simulations.
Validation of automated driving with a test environment in the cloud
Yongwook Lee CEO Elektrobit Automotive Korea Ltd Korea
Validating automated driving systems constitutes a major challenge because of their complexity, the need for decreasing development cycles, and strict legal safety requirements. In this talk, we show our approach to creating a toolchain for successful validation of such systems. As a basis, suitable ways of recording real driving scenes and generating simulated driving scenes need to be in place. Next, we show how to bring the collected data to the development environment. Finally, leveraging the cloud allows for parallel test and nearly infinite scalability, making it possible to validate automated driving functions quickly and efficiently.
Autonomous driving technology, which is considered to be the core technology of ADAS and various driving assistance technologies (ADAS) that are rapidly spreading, is focused on technology development not only in automobile companies but also IT companies such as Google and Apple. Implementation technology to achieve full autonomous driving is also important, but verification techniques to ensure the safety of autonomous driving through several fatal accidents caused by autonomous driving cars are also becoming important issues with regulations. This paper introduces the verification method for verifying the new technical factors of autonomous driving technology by looking at the main features of autonomous driving technology from the viewpoint of verification. The presentation will offer a safety analysis perspective to secure the functional safety of the control logic in response to various failure situations, taking into account the possibility of erroneous operation of the sensor-deep learning model implied by ADAS and autonomous navigation technology, which depend on machine learning and sensor technology. It will also explain how to deal with it. In addition, it will provide diverse post-analysis through in-vehicle/external data collection during actual vehicle driving, and suggest ways to supplement the running test that is absolutely insufficient through data-based simulation.
Hancom MDS presents developments in simulation, verification, prototyping for autonomous driving and ADAS 15:15 - 16:15
Dr EunDok Lee Senior researcher KATRI Korea
HIL verification based on camera, radar and ultrasonic waves for ADAS and autonomous driving
Hyeoul Woo Park Manager Hancom MDS Korea
Research institutes in the automotive industry are actively studying advanced driver assistance systems (ADAS) and autonomous driving that provide safety and convenience to drivers. In order to realize these benefits, various kinds of sensors are used. This presentation outlines solutions that validate ADAS technology using sensor simulation techniques during development.
Driving simulator for ADAS/AD (autonomous driving) controller testing
Hyunggu Kim Manager Hancom MDS Korea
Evaluation of the driver's comfort and safety awareness during the development of the ADAS and AD systems is essential. However, the part about human emotion is difficult to develop and verify because it is impossible to model. The presentation will introduce Cruden's driver-in-the-loop (DIL) simulation and a case study as a way to solve the problem.
MicroAutoBox embedded SPU (sensor processing unit) for prototyping complex perception algorithms on in-vehicle platforms
Seul Lee Manager Hancom MDS Korea
The presentation will outline the development of a vehicle sensor algorithm working with radar, lidar and camera, which is a key part of ADAS/autonomous ECU development. The speaker will introduce sensor algorithm development using MicroAutoBox embedded SPU in a vehicle environment, and sensor algorithm verification using a HIL simulator in an indoor environment.
Autonomous driving situation recognition and collision avoidance based on deep learning
Hyejin Kim General manager Hancom MDS Korea
The presentation will introduce a plan for a simulation test system for functional safety, after developing ADAS functions such as a vehicle collision avoidance system using an NVIDIA drive product based on deep learning vehicle image recognition and a vehicle CAN.
Please Note: This conference programme may be subject to change