HONDA The Power of Dreams

Reproducing the Driving Experience to Verify Functionality

As the name suggests, driving simulators are able to reproduce the driving experience. They replicate the feeling of driving to enable various functional verifications, essential for car development, to be conducted as if using an actual car.

In addition to being weather-dependent, verifications using actual cars can take a lot of time to swap out car parts and controls. Conducting verifications of automated driving (AD) and advanced driver assistance systems (ADAS) can also require testing in dangerous driving scenarios. And if multiple cars are required for the verifications, it can be quite difficult to ensure conditions are the same.

Driving simulators, on the other hand, allow testing of complex traffic environments under the same conditions every time, so that scenarios that are dangerous with actual cars can be tested safely. They even enable testing under special conditions that may not be possible in the real world. Additionally, simulators improve the efficiency of testing by dramatically reducing the time needed for rearranging parts, switching controls, and switching between test scenarios, all of which can take time with an actual car. At the end of the day, testing must be conducted on actual cars, but by using simulators as well, the key verifications can be completed in advance, with issues identified and addressed beforehand. As a result, they achieve faster development at lower costs.

Driving simulators are also effective tools for confirming feasibility of functions in the early stages of mass production development and when developing new technologies. Instead of relying only on desktop studies of value and effect, they are leading to new value creation because the verifications are achieved through real-life experience.

Honda uses driving simulators not only as tools to replace testing on actual cars, but as tools to understand human behavior as well. The aim is to study human characteristics through observation, measurement, and analysis of driver operations in the driving simulators, and then to create value that appeals to customers by leveraging the knowledge acquired there.

Honda maintains a number of large and small driving simulators for different purposes at its R&D facilities. This article provides information on some of those simulators together with examples of their use.

Driving Simulators for Evaluating Automated Driving and Advanced Driver-Assistance Technologies

There are two types of driving simulators for evaluating automated driving and advanced driver-assistance technologies; those focused on vision and those focused on the feeling of driving. Both types are used for studying concepts, analyzing requirements, and deciding specifications for both AD and ADAS systems. During testing, people drive the simulators and evaluate the acceptability and effectiveness of notifications they hear, in terms of ease of understanding and annoyance, and verify their effect in dangerous situations. At the same time, basic data on things like driving styles can be collected.

Actual cars can be installed directly into vision-focused driving simulators. In this example, a screen surrounding the car, with a radius of 5.5 meters and height of 8 meters, provides a wide field of view and long sight distance^*1 that ensure positioning and distance of landscape elements and physical objects around the car appear the same as in real life when sitting in the driver’s seat. The highly realistic vision encourages natural driving behavior.

This driving simulator provides particularly high-resolution images at the front of the car to improve visibility of traffic signals, signs, and other things in the distance. Images shown in the side and rearview mirrors reproduce actual the driving scenarios in real time. To prevent motion sickness, the vehicle is installed on a movable platform equipped with electric actuators that move the car in sync with the changing field of vision, which prevents any unnatural feeling and improves the immersive experience. In addition to offering a range of driving environments, from urban and suburban settings to highway driving, the simulator is also able to reproduce differences in weather and visibility conditions.

One example of a verification using this type of vision-focused driving simulator is to test an assumed system shutdown when using driving support functions, equivalent to automated driving level 2^*2, on a highway. The aim of the verification is to understand how the driver reacts in such a situation, and by how much the car deviates from its lane even if the driver reacts appropriately. Such knowledge is then used to develop appropriate specifications for commercialization of products. Members of the public participate at times to ensure the verifications capture realistic data from the market.

• 
• 
• 
• 
• 
• 

Feeling-focused driving simulators reproduce the forces (acceleration and deceleration) that the driver’s body is subjected to when accelerating and decelerating, and that reproduction is more realistic than vision-focused driving simulators. Like vision-focused driving simulators, these simulators can also use actual cars (a cut-body without a front and rear end). However, whereas vision-focused driving simulators install a car on a platform equipped with electric actuators, feeling-focused driving simulators install a cut-body within a shell-shaped dome, with height of 4.8 meters and width of 6 meters in this example, and the entire dome is moved by electric actuators. The cut-body is also moved by an electric yaw table within the dome to reproduce the translational motion (longitudinal, lateral, and vertical) and rotational motion (roll, pitch, and yaw) required for reproducing car behavior when driving.

With 360-degree projection of images across the interior surface of the dome, the rearview images as seen through the mirrors appear correctly. Stereophonic sound is also provided within the dome to reproduce sound direction, distance, and spread, thereby creating a realistic experience of things like other cars passing in adjacent lanes.

Feeling-focused driving simulators are more responsive than vision-focused driving simulators, which makes them more suitable for verifications of drivers’ responses requiring sudden braking or emergency maneuvers. One example is verifying driver reactions to a car, in an adjacent lane on a highway, suddenly cutting into your lane during traffic congestion. In automated driving level 3^*3, the system takes over from the driver to control driving operations, so the system must provide driving support that is as smooth as, or smoother than, human drivers. So the question becomes “how well do humans drive?” Therefore, one of the roles of this simulator is to collect basic data that would allow appropriate performance targets to be set for a practical application of automated driving level 3 technologies. As with vision-focused driving simulators, members of the public participate at times to ensure the verifications collect realistic data from the market.

Conducting such verifications when driving in the real world can be dangerous, so driving simulators can be used to ensure that testing is done safely. Equally, matching vehicle speeds and vehicle-to-vehicle distances is difficult to achieve in real-world driving, but simulators make the testing more efficient because they can reproduce the same conditions every time. Driving simulators partially replace the real world with virtual reality, measuring the driver’s senses and providing feedback for development. In this way, they closely relate to Honda’s Three Realities Principle for making decisions based on ‘going to the actual place,’ ‘knowing the actual situation,’ and ‘being realistic.’

• 
• 
• 
• 
• 
• 

*1 The distance from the driver’s eyes to an object. At distances of 5 meters or more, it feels almost like driving an actual car.
*2 At this level, driving operations are performed by a human. The system only provides driving support.
*3 At this level, driving operations are performed by the system. In limited areas that meet specific driving environmental conditions, such as during traffic congestion on highways, the system is able to take over from the driver to control driving operations while monitoring the surrounding traffic conditions. When conditions deviate from those required for system use, the system alerts the driver, who must take over vehicle handling immediately.

Small Driving Simulators for Evaluating Automated Driving and Advanced Driver-Assistance Technologies

Like vision- and feeling-focused driving simulators, small driving simulators are able to conduct functional verifications of AD/ADAS systems, but they are used when more agile verifications are required. When many conditions must be met, the tendency is for the simulators to become larger, but that means more time is required for verification setup.

Ease of operation is a characteristic of small driving simulators. They still have environments that create a sufficiently immersive experience, with coordination of images, sounds, and movements. However, they are able to perform verifications faster because reduced performance means less time needed for setup compared to simulators that can use actual cars. Another characteristic is that they use devices, including steering wheels and instrument panels, that are designed for easy rearrangement. They are used when speed is needed, such as studying concepts in the early stages of development, when confirming directions, and when conducting final verifications prior to the start of production of each device and control specifications.

Like the other driving simulators, small driving simulators can also be used for taking measurements to understand human behavior. Examples of this include using gaze tracking data to confirm where the driver is looking while driving, measuring reaction times when a warning sounds, and measuring reactions when different control and notification methods are used. Human reactions also vary depending on the day and physical condition. Honda is measuring and collecting data on a range of human conditions by looking at eye movement, head shaking, hands sweating, heart rate, and other factors, which it then works to quantify, model, and use in development. In fact, these simulators are practical applications of the human-centered approach that Honda has followed continuously throughout its history.

As with the other driving simulators, Honda provides opportunities for members of the public to use these small simulators as well to ensure collection of realistic data from the market. They are small and maneuverable enough to fit in a conference room, and they can be moved to other locations as needed.

Taking advantage of their location, close to the offices developing AD/ADAS systems, small driving simulators are being used for such things as quickly testing new ideas in the concept stage and discussing those ideas among staff members from relevant departments. This is because some ideas can only emerge when actually using something like this, rather than in desktop studies. Honda has a culture of “Waigaya” (open and frank discussions) that it uses to generate new value, with associates engaging in full debate, regardless of age or position. Small driving simulators provide development teams with environments for engaging in rapid cycles of checking effects and identifying issues through experience-based “Waigaya.” Honda also conducts ongoing studies of human behavior using the data acquired through these driving simulators to provide its customers with new value unique to Honda.

• 
• 
• 
• 

Driving Simulators for Evaluating Automobile Dynamic Performance

There are two specific applications for driving simulators for evaluating automobile dynamic performance. The first application, during mass production development, is for conducting sensory evaluation and safety verification of dynamic performance prior to testing on actual cars. At the end of the day, testing is conducted on actual cars, but using driving simulators in this stage aims to improve readiness of the technology and efficiency of the development process. As much as possible before testing on actual cars, verifications and confirmations are conducted, issues are identified, and solutions are implemented.

The second application is for considering future technologies and product value without actually developing a car. The aim of using these driving simulators for verifications, instead of actual cars, is to provide customers with new value and appealing products at lower costs and in shorter times.

Driving simulators for evaluating automobile dynamic performance use a cut-body, which is an actual car without the front and rear end, supported by six-axis electric actuators. The actuators are connected to a platform that floats on a cushion of air, blown at the metal table from the platform, in the same way that an air hockey table operates. Three-axis actuators are then used for sliding the platform. The main role of the six-axis electric actuators is to reproduce high-frequency movements such as when starting off and road surface inputs. The main role of the three-axis actuators is to reproduce low-frequency movements such as the continuous g-forces felt during braking, acceleration, and turning. In this setup, a boarding bridge electrically extends to a vehicle situated at a high level, with the relationship between the boarding bridge and vehicle heights adjusted to maintain an access height equivalent to getting into an actual car. This aims to prevent any unnatural feeling associated with driving the simulator, and feels just like getting into an actual car.

These simulators occupy a large work area, but that is still not enough to fully cover all the car movements to be reproduced. An important technology for this purpose is cueing, which is used to reproduce movements and create behavior that feels the same as driving an actual car, even though the simulator displacements are smaller than actual displacements. Honda has developed its own cueing technology that faithfully reproduces the movements necessary for evaluating dynamic performance while realizing lag-free vehicle behavior in response to driver inputs.

Driving simulators for evaluating automobile dynamic performance enable verification testing of handling stability, ride comfort, and AD/ADAS technologies under the same conditions as experienced in real car tests at Honda’s Tochigi Proving Ground and Takasu Proving Ground. Having measured and converted the road surfaces at each proving ground into digital data, the simulators reproduce a model car in real time with the same systems as an actual car. Honda’s R&D facilities also maintain bench testing equipment that measures and digitizes tire characteristics and suspension geometries. The model car incorporates data from this equipment as well.

Because the model car is reproduced with the same systems as an actual car, the person driving the simulator feels the exact same road surface inputs, via the tires and suspension system, as the actual car. Vehicle behavior in response to steering wheel operation is also the same. Controls and vehicle parts that could take months to replace and verify safety, can be changed and replaced by simply flipping a switch in the control room, so these simulators can dramatically reduce development lead times.

Like the other driving simulators, driving simulators for evaluating automobile dynamic performance can reproduce special test environments that may not be possible in the real world. For this reason, they are used as tools for understanding human characteristics. For example, they are used to observe, measure, and analyze driver operations and reactions by having the simulator driver maintain a target line displayed on the screen. Some areas of the driver’s sensory experience remain unknown and cannot be completely quantified. Honda is working to quantify these characteristics through a range of verifications. It is conducting such research and development to realize the philosophy of “the joy and freedom of mobility” that it has followed throughout its development of cars, and to achieve the feeling of a car that operates “at the will of the driver.”

• 
• 
• 
• 
• 
• 

Riding simulator for evaluating motorcycle dynamic performance

Motorcycle riding simulator for evaluating dynamic performance also being developed