jump to navigation

Autonomous Driving Vehicle Network December 10, 2012

Posted by rzmuedu in Mobility.
trackback

I live about 60 minutes away from work and commute on highway every day. I have always been talking to friends hoping that cars can drive themselves safely and efficiently. I’m taking this opportunity to design my version of autonomous driving system.

Autonomous driving really includes two big pieces, the car itself and the driving network. The reason that the system involves two big pieces is that having one autonomous driving car doesn’t solve the problem because still too many variables exist in the system and over complicate the driving system. For example, your self-driving car is cruising on the highway. All of a sudden another car next to you cut you off. Even if we are sitting behind the wheel driving will have a hard time coping with situation like that and I believe a good shaped human neuron system reacts faster than the complicated self-driving system. In a situation like that, the self-driving system will encounter a sudden change from many inputs such as range sensing, camera view, etc. The system will break hard based on the speed and range combination and it won’t be a happy ride for the passenger. However, if all cars on the road are self-driving and they can talk to each other, not only we won’t have the sudden cut line behavior, but the traffic will run so much smoother. Just like what we saw in the “I, Robot” movie. We all agree that Will Smith’s driving behavior is dangerous and not acceptable.

Extensive experiments had been done on automatic driving system such as Google who recently got their permit to have a self-driving car road test in Nevada. Therefore it is a proven concept that a self-driving car is not a dream. Although from what I learned, usually these cars run at a lower speed just to be safe because higher speed means all computation and reaction must happen in a shorter time. Given the complexity of the system, it will be very challenging.

In a high level design, the autonomous driving vehicle should include the following sub-systems. Range sensing, steering control, break control, acceleration control, visual sensing, and a “brain”. It is a mimic of human being. The range and visual sensing are derived from our eyes that can see what’s going on around the vehicle and determine how far away we are from an object. Our eyes and brain works together to derive more information such as speed and time to impact. Then we have our arms and hands to take care of the steering and feet to take care of break and acceleration. The self-driving car’s subsystems will function just like a human being to control the car.

The range sensing system needs to sense the front, back, both sides and beneath the car and constantly provides these information to the control system. The front and back range sensing helps to determine if the car is running into something when going forward or backward respectively.  The back range sensing can also act as rear end avoidance. For example a car is running into you from behind and the back range sensor picked that information. The control system can drive the car out of the way to avoid the collision which sometimes can also cause fatal injury to passenger. The side range sensing is used guide lane change. The sensor sensing the beneath information can help avoid driving over something that’s too high and can cause damage to the drivetrain underneath the body.

The steering, breaking, acceleration control subsystem can take care of their functions by executing commands sent from the control system. These three system will work together to avoid any dangerous driving behavior such as making sharp turn at high speed or anything that can cause damage to the car or passenger. These are rules and restriction implemented in the control system.

The visual subsystem is very important. It can contain not only optical sensors but also IR sensors to pick up visual information when the lighting condition is bad such as driving at night/fog/rain. These sensors will pick up 360 degree visual information to provide complete view to the control system. Combined with good image process software, recognition software, the control system can then pick up useful visual information such as road signs, line guide, and much other good information. However, image processing is computation heavy operations that require time and power. Proper balance of the software must be taken into consideration. With image recognition, the system can pick up landmarks to better locate itself along with the help of GPS. This is especially important in big cities since the GPS signal is either blocked or bounced off accuracy by all the tall buildings and leaving it very unreliable.

Once the self-driving system is completed and proven to be working, the network between cars can be established. This means some more components to be added to the car.

One network is the GPS network. Utilizing the GPS satellites, the car can get a very good idea of where it is when in open field. Then using digital map information such as Google Map, the car can place itself and plan route to the destination. After that, cellular network can also come into play to provide additional traffic, weather information. With the development in cellular network, it is very easy to transmit data to the control system to let it know if there is traffic jam on the route or road construction or bad weather in front. The control system can take all information into consideration and plan a better route based on the user’s specification of either faster time or shorter distance. Next is the ad-hoc network between self-driving cars. A good wireless communication between cars nearby can form an ad-hoc network. Then they can start sharing useful information such as travelling speed, destination, current state, next move, etc. This is very useful in eliminating some of the uncertainty in the overall environment since the car’s nearby neighbors has shared all their information with you. For example, the car on your left needs to get off the highway on the next exit and needs to move to your lane. Your car which is driving at a constant speed maintaining a safe distance from the car in front of you, can then slow down a little to create enough space for the car exiting the highway to complete their lane change move. Since everything is known, there is no surprise like cutting lane, the passenger will experience a smoother ride and less traffic jam is likely to form.

Benefits from the system are obvious, less human involvement means less chance of human error. How many times did we hear someone stepped on the gas when they meant to break? Missing a highway exit will never happen. Reckless driving disappeared from the road. Higher fuel efficiency and less traffic jam. The benefits are countless.

Of course testing such a system needs to be comprehensive and constantly done. Besides cars, there are pedestrian, bikers, roller-skates on the street. They don’t have the deterministic equipment installed and can’t share their information with the autonomous driving cars. They are the variables in the system now and any damage to them by the autonomous driving cars system will be fatal and put the system in danger. Regulations need to be modified in order for it to come true, as well as mandate of only autonomous cars can go on street.

Bibliography

Kageyama, Y. (2012, 11 30). Toyota Smart-Car Technology Let Autos ‘Talk’ To Each Other, Sense Pedestrians And Red Lights. Retrieved from Huff Post Tech: http://www.huffingtonpost.com/2012/11/12/toyota-smart-car-technology_n_2115570.html?utm_hp_ref=technology

Ma, J. (n.d.). Ford, GM, Tech Firms Driving To ‘Smart’ Car Era. Retrieved from Investor.com: http://news.investors.com/technology/040711-568445-ford-gm-tech-firms-driving-to-smart-car-era.htm#ixzz2C1RCRWVZ\” data-mce-href=

Newcomb, D. (n.d.). You won’t need a driver’s license by 2040. Retrieved from CNN, Wired: http://www.cnn.com/2012/09/18/tech/innovation/ieee-2040-cars/index.html

Advertisements

Comments»

No comments yet — be the first.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: