After the recent article about yet another Model 3 driver asleep on AP and apparently using a weight on the steering wheel to fool the steering torque, it is making me think more and more that Tesla really should consider using the in-cabin camera in the Model 3 to monitor driver attention and allow hands free driving. Now, I am aware that Musk is against the idea because he wants to focus on just getting to L4/5 autonomy where driver attention is irrelevant. So he sees a driver attention system using the camera as a waste of time and resources. But I think there are good reasons for and against the system. I outline them below.
Some main reasons FOR:
1) No major cost in time or resources
The in-cabin camera is already there. So no new hardware is necessary. Tesla would just need to do the software part and push out the feature as an over the air update. And the software for facial recognition is pretty common place now. It would not be a huge cost in time or resources.
2) Increased safety
Adding a driver attention system using the camera would increase safety since it would prevent people from cheating the nag system. It would also increase safety by providing an additional back up since you would have two systems, the torque and the camera to make sure the driver is alert. And we've certainly seen several instances where the torque system alone was not good enough. It could be argued that the only way to make sure that a driver is indeed paying attention to the road is by directly monitoring their face.
3) It would allow true hands-free driving in some instances
Like in other cars, using the camera to monitor the driver would allow for true hands-free driving in some instances. Tesla could design the feature to only activate hands-free driving in certain situations where AP or NOA have a high confidence such as routine driving on the highway with no or little traffic. And the feature could be designed to turn off hands-free driving when confidence is not high enough, such as city streets, difficult lane changes, on ramp off ramp maneuvers, busy traffic. This would allow AP to be a L3 autonomous system that would allow for true hands free driving when it is safe to do so. Plus, Tesla could gradually expand when hands-free driving is allowed as they improve AP towards FSD. This would be really nice for drivers I think.
4) It would serve as a nice stop gap until we get to FSD
Sure, eventually with FSD, monitoring driver attention will become unnecessary. But we are not there yet. So why not add this feature now to give owners hands-free driving and increased safety while we wait for FSD?
Some main reasons AGAINST:
1) The in-cabin camera may not be good enough to implement the system properly since it is not a IR camera and is not positioned in front of the driver. If true, then implementing the feature could prove difficult or flawed. In which case, it might not be worth it.
2) Only the Model 3 right now has the in-cabin camera which means that only the Model 3 would be able to get this feature. As a Model 3 owner, I admit that I am biased since I would benefit from the feature. I admit that it would put the Model S and X at a disadvantage. It would certainly suck for Model S and X owners that they can't get true hands-free driving even though they drive the more expensive car. Of course, Tesla could add an in-cabin camera for monitoring driver attention in the Model S/X refresh to fix this issue. That would even give the S and X refresh an additional demand bump.
3) Like Musk's argument, it could be argued that it is better to just focus on making AP more FSD and safer rather than spending resources on adding a driver attention system that ultimately will be unnecessary. After all, the sooner Tesla gets to FSD, the better.
4) There are better ways that Tesla could address the problem of cheats than to devote resources to a camera based driver monitoring system. For example, Tesla could probably just change the torque requirement via software update so that cheats like autopilot buddy don't work anymore.
Thoughts?
Some main reasons FOR:
1) No major cost in time or resources
The in-cabin camera is already there. So no new hardware is necessary. Tesla would just need to do the software part and push out the feature as an over the air update. And the software for facial recognition is pretty common place now. It would not be a huge cost in time or resources.
2) Increased safety
Adding a driver attention system using the camera would increase safety since it would prevent people from cheating the nag system. It would also increase safety by providing an additional back up since you would have two systems, the torque and the camera to make sure the driver is alert. And we've certainly seen several instances where the torque system alone was not good enough. It could be argued that the only way to make sure that a driver is indeed paying attention to the road is by directly monitoring their face.
3) It would allow true hands-free driving in some instances
Like in other cars, using the camera to monitor the driver would allow for true hands-free driving in some instances. Tesla could design the feature to only activate hands-free driving in certain situations where AP or NOA have a high confidence such as routine driving on the highway with no or little traffic. And the feature could be designed to turn off hands-free driving when confidence is not high enough, such as city streets, difficult lane changes, on ramp off ramp maneuvers, busy traffic. This would allow AP to be a L3 autonomous system that would allow for true hands free driving when it is safe to do so. Plus, Tesla could gradually expand when hands-free driving is allowed as they improve AP towards FSD. This would be really nice for drivers I think.
4) It would serve as a nice stop gap until we get to FSD
Sure, eventually with FSD, monitoring driver attention will become unnecessary. But we are not there yet. So why not add this feature now to give owners hands-free driving and increased safety while we wait for FSD?
Some main reasons AGAINST:
1) The in-cabin camera may not be good enough to implement the system properly since it is not a IR camera and is not positioned in front of the driver. If true, then implementing the feature could prove difficult or flawed. In which case, it might not be worth it.
2) Only the Model 3 right now has the in-cabin camera which means that only the Model 3 would be able to get this feature. As a Model 3 owner, I admit that I am biased since I would benefit from the feature. I admit that it would put the Model S and X at a disadvantage. It would certainly suck for Model S and X owners that they can't get true hands-free driving even though they drive the more expensive car. Of course, Tesla could add an in-cabin camera for monitoring driver attention in the Model S/X refresh to fix this issue. That would even give the S and X refresh an additional demand bump.
3) Like Musk's argument, it could be argued that it is better to just focus on making AP more FSD and safer rather than spending resources on adding a driver attention system that ultimately will be unnecessary. After all, the sooner Tesla gets to FSD, the better.
4) There are better ways that Tesla could address the problem of cheats than to devote resources to a camera based driver monitoring system. For example, Tesla could probably just change the torque requirement via software update so that cheats like autopilot buddy don't work anymore.
Thoughts?
Last edited:
