To really test the ideas against each other you need to gear up the drive, so that you actually CAN break traction on the wheels while cornering. You are right about the observation that the lower performance of speed controlled vehicle is due to the control reserve, but also very right that you can't really give up that margin while retaining speed control. Using torque vectoring effectively gives you that edge just because it's stable at 100% duty cycle on the outside wheel. If you manage to get that vehicle fast enough so that it slides out in the corners, you are likely to observe a bigger edge towards torque vectoring, as it will keep the cornering forces of the front wheels useful. Thanks for the update! You may mention next time, where the input came from 😛
I tuned the Torque Vectoring algorithm to edge of understeer. Interesting fact: with Ackerman I need less power reduction on the inside wheel (-40%) than with a parallel rack (-60%). So Ackerman not only makes tighter turns, it also gets more out of the cornering wheels! It is really difficult to make LEGO wheels break traction. I've only had success with rubberless wheels and rotocasters.
@@AntonsMindstorms you may try with the old technic wheels with integrated tyres, the ones from the universal sets, that have round axle holes (so you need the belt wheels put inside the rims to drive them). They have quite poor traction, but still better than blank plastic. I'll have to dig those up and do some tests. Btw: to check traction in corners you'd usually keep the steering angle constant and keep increasing speed until you break traction. You should be able to actually see the limit approaching by watching the gyro signal, but flying out of the corner is just as good an indicator - at least as long, as toy cars are concerned 🤪
@@MarekLewandowski_EE High-powered RC cars use a diff gear and a single motor and a gyro for drift control. It also looks cool, but I guess it's different because the gyro widgets seem to change your steering angle. ua-cam.com/video/sx1CNSHLc10/v-deo.htmlsi=4s7SkQmFkubJeaxV&t=248
@@AntonsMindstorms that's a different pair of socks, that's a sort of a mistuned ABS/ESP unit. What I mean is, when you circle faster and faster, at first your raw hub gyro signal should show a constant, slowly increasing yaw rate, corresponding to your angular speed around the loop. As you approach the point of loss of traction this signal may start having some unsteadiness to it, as the wheels start to slip on bumps but catch on flats. I haven't tried it with Spike hub yet, so I'm not sure if this is going to be visible
This guy is a genius
To really test the ideas against each other you need to gear up the drive, so that you actually CAN break traction on the wheels while cornering.
You are right about the observation that the lower performance of speed controlled vehicle is due to the control reserve, but also very right that you can't really give up that margin while retaining speed control. Using torque vectoring effectively gives you that edge just because it's stable at 100% duty cycle on the outside wheel. If you manage to get that vehicle fast enough so that it slides out in the corners, you are likely to observe a bigger edge towards torque vectoring, as it will keep the cornering forces of the front wheels useful. Thanks for the update! You may mention next time, where the input came from 😛
I tuned the Torque Vectoring algorithm to edge of understeer. Interesting fact: with Ackerman I need less power reduction on the inside wheel (-40%) than with a parallel rack (-60%). So Ackerman not only makes tighter turns, it also gets more out of the cornering wheels! It is really difficult to make LEGO wheels break traction. I've only had success with rubberless wheels and rotocasters.
@@AntonsMindstorms you may try with the old technic wheels with integrated tyres, the ones from the universal sets, that have round axle holes (so you need the belt wheels put inside the rims to drive them). They have quite poor traction, but still better than blank plastic. I'll have to dig those up and do some tests.
Btw: to check traction in corners you'd usually keep the steering angle constant and keep increasing speed until you break traction. You should be able to actually see the limit approaching by watching the gyro signal, but flying out of the corner is just as good an indicator - at least as long, as toy cars are concerned 🤪
@@MarekLewandowski_EE High-powered RC cars use a diff gear and a single motor and a gyro for drift control. It also looks cool, but I guess it's different because the gyro widgets seem to change your steering angle. ua-cam.com/video/sx1CNSHLc10/v-deo.htmlsi=4s7SkQmFkubJeaxV&t=248
@@AntonsMindstorms that's a different pair of socks, that's a sort of a mistuned ABS/ESP unit. What I mean is, when you circle faster and faster, at first your raw hub gyro signal should show a constant, slowly increasing yaw rate, corresponding to your angular speed around the loop. As you approach the point of loss of traction this signal may start having some unsteadiness to it, as the wheels start to slip on bumps but catch on flats. I haven't tried it with Spike hub yet, so I'm not sure if this is going to be visible