3 ms·

Curious whether they could take this even further. I've always wanted to try driving where turning uses absolute position rather than relative motion on the st

by derefr 5y ago

Curious whether they could take this even further. I've always wanted to try driving where turning uses absolute position rather than relative motion on the steering column to determine wheel rotation, like an analog stick in a driving game: rotate fully to the right and the gear box will push the wheels continuously to the right; let go and the yoke would snap back to neutral, and the gear box would then work with sensors to bring the wheels to a straight-forward neutral position. The absolute yaw of the yoke at any moment should be reflected in the absolute yaw of the wheels. I'd also be curious, given a flight yoke, to try controlling acceleration (not velocity!) using the Z dimension of the yoke, like actual planes do — maybe not with actual travel, but just with strain gauges on the shaft of the yoke to detect how hard you're pushing/pulling on it. Push to accelerate; pull to decelerate; relax to cruise.

systemvoltage 5y ago
That sounds tiring and painful. Heel of your foot helps to stabilize the throttle actuation. Your hands would be hanging in the air. Also there is this - accelerate accidentally while also turning. There is absolutely zero need to reinvent the car UI/UX. Don’t try to fix what’s not broken.
- nitrogen 5y ago
  There is absolutely zero need to reinvent the car UI/UX. Don’t try to fix what’s not broken. While nobody has yet improved significantly on the traditional control layout of gas/brake/[clutch]/wheel, I still think as a matter of principle that it's really unlikely we are at the global optimum. Thus, experimentation should always be on the table, as long as it's optional.
  - systemvoltage 5y ago
    Even if we're not at global optimum, we've got some billions of users that are used to the current UX/UI.
- derefr 5y ago
  We were talking about "driving", but who said anything about cars? ;) "Giving strategic-level input to the AI that drives around the self-driving vehicle that you have final responsibility over" doesn't necessarily have much to do with "driving a car." "Being the remote operator who steps in to remote-drive a self-driving vehicle when the AI can't figure out what to do" has even less to do with "driving a car." (Presuming they don't give each remote operator a full 360° driving simulator, it's only going to be about as immersive as a video game, anyway.) IMHO, it's only natural to begin re-thinking "car" UI/UX for those cases. Especially for the future generations who will be born into an era where they may "learn to drive" and then "drive" all their lives, never using any car other than self-driving ones. It's only during this transitional era, where autonomous driver AI isn't very good and needs constant manual input — and where the majority of people who own self-driving cars have also owned "hand-driven" cars — that it would seem to be a good idea to give input to a driving AI using hand-driven car inputs.
nitrogen 5y ago
Push to accelerate; pull to decelerate; relax to cruise. I've had similar ideas for speed control, but with something in the gear/handbrake area. The problem I couldn't solve was choosing an orientation and response curve that didn't make it so acceleration made it hard to brake, and braking didn't make it hard to change the amount of braking, while still feeling cool (i.e. not left/right oriented).
- derefr 5y ago
  My thought would be that "choosing to accelerate [more/less]" and "choosing to brake [more/less]" are strategic-level decisions; and, if you have a self-driving car, the tactical-level fine input of how much more/less can be handled for you. Kind of like what's going on with mouse-"controlled" movement in this AI gameplay here: https://www.youtube.com/watch?v=qYluZRwrw9w https://www.youtube.com/watch?v=qYluZRwrw9w Or, to put that another way: the car already has a (low-confidence) idea that there's an obstacle in the path, and so it already has a "possible plan" to brake hard to stop in time to not collide with the obstacle. You just have to choose that plan — declare your strategy — by pulling back on the yoke. It then executes on the plan.
landemva 5y ago
Joystick steering, with thumb and trigger controls. Good times to be had! https://www.komatsuamerica.com/equipment/wheelloaders/large/wa600-8 https://www.komatsuamerica.com/equipment/wheelloaders/large/...

by derefr 5y ago

systemvoltage 5y ago

That sounds tiring and painful. Heel of your foot helps to stabilize the throttle actuation. Your hands would be hanging in the air. Also there is this - accelerate accidentally while also turning. There is absolutely zero need to reinvent the car UI/UX. Don’t try to fix what’s not broken.

nitrogen 5y ago
There is absolutely zero need to reinvent the car UI/UX. Don’t try to fix what’s not broken. While nobody has yet improved significantly on the traditional control layout of gas/brake/[clutch]/wheel, I still think as a matter of principle that it's really unlikely we are at the global optimum. Thus, experimentation should always be on the table, as long as it's optional.
- systemvoltage 5y ago
  Even if we're not at global optimum, we've got some billions of users that are used to the current UX/UI.
derefr 5y ago
We were talking about "driving", but who said anything about cars? ;) "Giving strategic-level input to the AI that drives around the self-driving vehicle that you have final responsibility over" doesn't necessarily have much to do with "driving a car." "Being the remote operator who steps in to remote-drive a self-driving vehicle when the AI can't figure out what to do" has even less to do with "driving a car." (Presuming they don't give each remote operator a full 360° driving simulator, it's only going to be about as immersive as a video game, anyway.) IMHO, it's only natural to begin re-thinking "car" UI/UX for those cases. Especially for the future generations who will be born into an era where they may "learn to drive" and then "drive" all their lives, never using any car other than self-driving ones. It's only during this transitional era, where autonomous driver AI isn't very good and needs constant manual input — and where the majority of people who own self-driving cars have also owned "hand-driven" cars — that it would seem to be a good idea to give input to a driving AI using hand-driven car inputs.

nitrogen 5y ago

Push to accelerate; pull to decelerate; relax to cruise. I've had similar ideas for speed control, but with something in the gear/handbrake area. The problem I couldn't solve was choosing an orientation and response curve that didn't make it so acceleration made it hard to brake, and braking didn't make it hard to change the amount of braking, while still feeling cool (i.e. not left/right oriented).

derefr 5y ago
My thought would be that "choosing to accelerate [more/less]" and "choosing to brake [more/less]" are strategic-level decisions; and, if you have a self-driving car, the tactical-level fine input of how much more/less can be handled for you. Kind of like what's going on with mouse-"controlled" movement in this AI gameplay here: https://www.youtube.com/watch?v=qYluZRwrw9w https://www.youtube.com/watch?v=qYluZRwrw9w Or, to put that another way: the car already has a (low-confidence) idea that there's an obstacle in the path, and so it already has a "possible plan" to brake hard to stop in time to not collide with the obstacle. You just have to choose that plan — declare your strategy — by pulling back on the yoke. It then executes on the plan.

landemva 5y ago

Joystick steering, with thumb and trigger controls. Good times to be had! https://www.komatsuamerica.com/equipment/wheelloaders/large/wa600-8 https://www.komatsuamerica.com/equipment/wheelloaders/large/...