Mando’s self-driving demonstration, ‘Hockey’ proves to be a “smart and sophisticated" vehicle
January 16, 2019, on a 4-lane highway near the Pangyo Mando Global R & D Center. Mando's autonomous driving vehicle 'Hockey' was standing by for its demonstration.
Approved for autonomous driving by the Ministry of Land, Infrastructure, and Transport in 2017., ‘Hockey’ is a Black Genesis G80 vehicle armed with a vast array of advanced technologies developed by Mando.
An autonomous driving vehicle refers to a car that uses sensors (e.g. radars, cameras, etc.), instead of relying on the driver’s eyes, to recognize its surroundings and control the steering wheel, the brake and accelerator pedals.
Therefore, as an autonomous driving vehicle, 'Hockey' is equipped with a myriad of high-performance sensors that recognize the external environment.
‘Hockey’s’ front bumper, rear bumper, and side mirrors have as many as 10 Lidars and cameras.
It has a large, cylindrical Lidar on the roof to enable more accurate positioning, and it also has a GPS receiver and a black box.
Using a suite of high-performance sensors, ‘Hockey’ supports a number of Advanced Driver Assistance Systems (ADAS) such as Adaptive Cruise Control (ACC), Autonomous Emergency Braking (AEB), and Lane Keeping Assist System (LKAS).
The first thing I noticed when I opened ‘Hockey’s’ door and settled down in the back seat was how overwhelming the interior was with a console box full of various electronic devices and monitors.
When the PC in the passenger seat came online, the monitors in the front of each seat started to run their assigned protocols.
It was obvious that ‘Hockey’ was not your average car.
'Hockey' required a few minutes to spool up its systems and warm all of its self-driving systems up.
Although ‘Hockey’ can obtain highly accurate positioning data using its GPS exclusively in open spaces, the Mando Global R&D Center was surrounded by buildings and built-up areas, which, according to an official from Mando, meant that it needed some time to integrate its GPS information with information from its 360-degree surround camera system.
As the official from Mando turned the ignition on and began driving, the monitor displayed coordinates acquired through GPS in green and coordinates acquired based on Mando’s algorithm in red. In other words, 'Hockey’s' actual position was displayed in red.
Once we arrived in a wide-open area, the GPS coordinates began to match the actual position of 'Hockey.' But, again, as the vehicle got closer to the Mando Global R&D Center, the two coordinates started to drift apart.
Surrounded by tall buildings in the city, GPS alone is not enough to pin point the position of a vehicle. Fortunately, ‘Hockey’ uses a high-precision map and positioning algorithm developed by Mando to calculate its exact position as well as the actual width of the road lane.
The Mando official explained that navigational applications using standard GPS data could be off by up to 5m, whereas ‘Hockey’ navigational system could be off by just 10cm.
I was genuinely thrilled to learn how sophisticated 'Hockey' was. Before the test drive, I opened the ‘Hockey’s’ boot to see if it had anything in there. It was filled with all sorts of advanced equipment, which I studied with great curiosity for a while.
Mando's autonomous driving level-4 is equivalent to the Full Self-Driving Automation level defined by the National Highway Traffic Safety Administration (NHTSA), which means that the car will travel to its destination without any input or intervention from its driver.
In other words, the person inside the vehicle can do other things (e.g. reading or playing games) while the vehicle does all the driving on its own.
The official from Mando did warn us that something unexpected could happen during the test drive since we were expected to drive on open roads, not on a test track according to a predetermined scenario.
As soon as 'Hockey' started up, the monitors came to life.
'Hockey' collected information with its six Lidar sensors, and displayed vehicles in the shape of rectangular boxes. Depending on the characteristics of each vehicle, the boxes appeared to be in different sizes and colors.
Cars in motion were blue, stationary cars were red, cars coming from the opposite direction were yellow, and cars moving in a diagonal direction were pink. The bigger boxes were either buses or trucks.
The position of 'Hockey,' the lanes, the expected route, intersections, other cars, and several other elements, were all easy to check and understand in real-time. It was easy to tell which vehicle was coming from behind, or whether it was in the overtaking line, without looking at the wing mirrors and rear-view mirror.
In particular, when ‘Hockey’ was switching lanes, I did not have to turn my head because there were no blind spots. I just had to look at the monitor.
The system constantly d the composition and position of its colored boxes, the circles, and solid lines. I almost felt as if I was looking at a computer game.
When we drove about 100m ahead of the Pangyo intersection, I was told that the upcoming traffic light will turn green by the time 'Hockey' reached the intersection. Surely enough, the traffic lights turned green and we were able to pass through without having to slow down.
Now, at intersection, there is a signal controller operated by the National Police Agency. By working with a telecommunication service provider affiliated with the NPA, 'Hockey' is aware of when a traffic light will turn red ahead of time. That was how it could calculate how much time it had before reaching a specific traffic light, and whether it could pass the intersection based on the remaining distance and speed of the vehicle. The more I watched ‘Hockey’ work, the more I realized the car was such a smart and sophisticated machine.
The Mando official said that in Pangyo, the roads have plenty of curves, randomly parked cars, and a lot of light reflected from building facades made of glass, making autonomous driving a difficult talk even in clear weather.
Even in these challenging conditions, ‘Hockey’ safely completed the 2.7km-route around the public road test section in Pangyo and returned to the starting line.
For the test drive, 'Hockey' used its cognitive technology, positioning technology, control technology, decision-making technology, and communication technology at a speed of approximately 40 km per hour completely without the need for manual intervention. ‘Hockey’ also changed lanes and negotiated with turns at intersections perfectly.
Overall, it successfully showcased just how advanced Mando’s self-driving technology is at the moment.
11For the test drive, Mando’s high-precision positioning system (with SVM) and route generation system (with model predictive control) were the key features. Mando showed that it had opened a new door to autonomous driving thanks to countless test runs and experiments using complex testing sections in the Pangyo area. Moving forward, Mando will incorporate AI and other technology in its autonomous driving vehicle ‘Hockey’ to improve its capacity to recognize surrounding conditions and predict future events.