![]() |
Remote Kill Switch
I was reading another thread "Safety Issue: Robots Moving in Pits" and ask now, why don't we have a clear remote kill switch?
A lot of teams don't know/forget that the space bar on the computer is an emergence kill switch. As a result (also coming from some unsafe practices) people are getting run over by 100 lb robots. People can't argue that part. The judges have there own kill switch during the game. Something that would be very intuitive would be like a 2 button detonator switch (Separate from any other controller). The first button has a guard, and would have to be activated to run the robot (means someone has to hold the kill switch). The second of course would be the kill. This takes a VERY proactive approach to having someone specifically watch for emergencies. No, you can't stop people from making bad choices, but you can make it easier to make good ones. I think I'll be bringing this up at our next team meeting. Thoughts? |
Re: Remote Kill Switch
That used to be required, before they switched to using the space bar as an emergency stop. It was a red button (connected over USB) that said stop on it, and it sent a keystroke (Alt+Enter, I think) to the driver station to emergency stop the robot.
Its in the bottom left of this picture: http://team358.org/files/programming...verStation.JPG |
Re: Remote Kill Switch
Quote:
Yes, keeping an eye on robot going rogue and stopping it is not easy. I think instead of making more rules and more "safety" agents patrolling the pits, every team should be proactive and train their robot operators. First thing is to stop the robot from randomly running around and crashing into someone or something (rogue robot?). Our team has been trained to put the robot on wood blocks and ensure that wheels don't touch them, even if we are not testing the drive. Other thing is about actuators/manipulators, each season and every team has different design. Easier and a common solution would be to have a "kill radius" as recommended by others. so make sure that everyone stands beyond this imaginary circle, even if you are not testing manipulators. This could reduces the risk of accidents, and again there is no 100% guarantee that accidents don't happen. |
Re: Remote Kill Switch
An even better safety approach is a 'dead man' switch. This requires a person to continuously hold a switch in the pressed position for the robot to operate. Of course you couple this with a policy that you never jerry rig the switch to be held other than by a human watching the robot.
Basically every push lawn mower has one of these on the handle, and every riding lawnmower has a seat switch that does the same thing (at least when the blades are spinning.) Update: Clarification: I agree, driving in the pits is not appropriate. My suggestion above is more for driving in the shop and could even apply to driving on a practice field, which is more danger prone than a real field. Seems like it would be a good idea to have a driver station setting that allows you to invert the space bar, and require that it be pressed when enabling and result in a kill if released while enabled. |
Re: Remote Kill Switch
As a person that routinely volunteers as CSA/FTAA please do not make it possible for your robot to drive in the pit.
If you want to put it up on blocks or something that's one thing. Even better if the venue can help you do that safely with materials. If you put it on the floor and try to drive I am sorry but I will object to that. Please use the practice fields to fully test your robot's work envelope. I am not against additional safety controls like the USB red button which is more recognizable to teams, or even requiring teams to put a label on the space bar in the rules (easy to do and cheap). However an emergency stop is not replacement for not creating an dire situation in the first place. |
Re: Remote Kill Switch
Quote:
|
Re: Remote Kill Switch
Quote:
A robot outside of transport configuration is probably larger. The totes are relatively large. I watched several people put their robot on the floor and test actuators and I am fine with that. However a large robot, not really designed with crumple zones or padding moving in a room full of tightly packed people? If you watch the Mount Olive District event I always loaded the Blue Alliance field first. That queue was slow to move. In walking by one of the robots on the field I got shirt caught by a piece of aluminum projecting only because the robot was in a perfectly reasonable configuration for the field. So realistically speaking, even if the robot is not driving, putting the robot fully expanded on the floor in the pits moving requires a fair deal of consideration to keep people out of the work envelope and the potential drop away of the tote. I wouldn't object as long as it was clear the area in the work envelope was actively patrolled. This could be easily addressed by the venue offering a floor level area for testing in addition to the practice field. For example at Mount Olive High School we certainly have a lot of floor we could offer for temporary use. If you walked back by our machine shop table there was a large section of the lobby back there empty. |
Re: Remote Kill Switch
Quote:
|
Re: Remote Kill Switch
Quote:
The dead man switch approach would be instituted in your shop/practice area at home (based on your perceived need for that level of safety) and could possibly be used at an event practice field where there are lots more people around often without field walls protecting them from bots. |
Re: Remote Kill Switch
A dead man switch would have to be implemented correctly, tested thoroughly and the users trained properly and thoroughly to ensure that it serves it's intended purpose. Otherwise, is a form of false security that creates yet more complexity with little value.
It is a fair bet that FIRST put a lot of time and effort into designing and performing validation testing on the E-Stop function integrated into the FMS. |
Re: Remote Kill Switch
In scenarios where we might need to stop the robot quickly, an operator stays by the driver station is trained to hit the spacebar if things get out of control. Hard to see how a dedicated button improves on that.
|
Re: Remote Kill Switch
Quote:
|
Re: Remote Kill Switch
The space bar & enter key work well enough. Keep in mind a real "ESTOP" would require interrupting the power pathways to the motors since one of the failure modes is a shorted H bridge or stuck Spike. Something not legal in current FRC rules. Just saying.
|
| All times are GMT -5. The time now is 01:36. |
Powered by vBulletin® Version 3.6.4
Copyright ©2000 - 2017, Jelsoft Enterprises Ltd.
Copyright © Chief Delphi