Lecture 10-9 Post-Mortem

The thing about this class is that, I probably know about 70%-80% of the content because I'm in computer science. Friends and teachers will sometimes show these nifty videos or Slashdot will mention it or it'll somehow cascade to you because it is relevant to what you do. It's something that the geek community passes around, revels, and applauds over.

So I suppose the reason why I am in this class is to see if there is something new to be absorbed, some different angle that was noted that no one has seen before. From a programmer's point of view, you see things in technical terms. (But before computers, I was really an avid artist in high school) And sometimes I look at things from an aesthetic point of view and I know why it didn't end up being successful despite the advanced technology it touts.

Therefore, what I do find interesting are the little side thoughts I have have when I review a lot of the information again.

For example, the car data extraction mentioned in class today (I had seen this earlier in a cognitive science class I had taken a few quarters back). Back then, I shrugged and said, "I guess it's interesting for the car makers to know this information". But this time. I noticed the 360 camera in the video and I sort of had a side thought of an idea:

The biggest concern with cars and driving in general are collisions.

Collisions happen when you don't see something, or see it but its too late to correct your mistake. As long as you are aware of everything around you and you don't actively collide into it, you should be fine (unless someone else runs into you).

Therefore, there should be something to consistently monitor what is around your car at all times.

Now I know in Japan, it's law to have the 4 corner cameras on your car, but that still means you have to monitor the objects in your car cameras.

But the idea I had was that you could create a detection system that monitored the space around your car. There can be a feedback either on a monitor that showed a representational model of your vehicle and all objects surrounding it or some other device to keep you aware of all things surrounding your vehicle at all times.

Now the system's advantage would be that if you were to turn your car, you would turn on your signal lights and it could scan the objects in the direction that you intended to turn your car. It would give a light 'be cautious' beep to tell you there is someone in the other lane. And a louder beep if there is someone in the lane within collision distance. (In addition, the system is also good because if forces drivers to turn on their turn signals -which is a huge problem and is also the cause of a lot of collisions today). This is a very initial idea and would need a lot of fleshing out, but perhaps it would be a better system than just putting camera monitors on the car's dash that the driver might not necessarily always be able to look at if they have to keep their eyes on the road.

Another thing I was interested in was the minicam that one of the researchers used, was it real time live stream or a series of intermittent pictures strung together? And if it's the former, it is certainly small enough to aid in surgical procedures. Why not use the tiny cams instead of the giant tubes you have to stick down a patient's throat?