# Is this kind of steering possible?

I would like to build a small robot and was wondering if a particular configuration of steering would even be possible and stable.

What I would like is to have the two rigid axles mounted on a pivoting point like the image. The pivoting points would be free to move on their axes without restrain. Steering should be achieved via the different speed of the 4 wheels (one motor on each wheel).

Problem is, if I start accelerating the outer wheels, to me it looks like it would start going in a diagonal direction, with both axles keeping parallel to each other rather than achieving something like the picture, even though the outer wheels on a corner should go faster than the inner wheels. Is this configuration even possible?

• you would need to either accelerate/decelerate opposite corner wheels. ie speed up front left and rear right, or decelerate front right and rear left. Dec 6, 2021 at 17:25
• That is what I also thought.. But, Aren't the inner wheels in a corner the slowest? Dec 6, 2021 at 17:39
• In theory it might be possible but you would have no constraints on the axles and if a wheel slips or spins that axle will turn violently. You would need servo motors to ensure that you could control the relative velocities accurately. Simple brushed DC motors would not suffice. You'd also need some sort of periodic calibration to correct any accumulated relative error between the two axles. If you did get it working the vehicle would turn about the point of intersection of the two axles. Dec 6, 2021 at 20:42
• If each wheel has its own motor, then you might as well vary their speed independently to properly follow the arch rather than having everything pivot. It will do the same thing. I mean, I guess with the pivot you can get a tigher turn radius without skid, but you'd have to decide if that was worth the effort. I imagine the control would be a bit wonky especially if the ground was loose and allowed skidding. You will need encoders on each motor at a minimum. Likely more sensors if on loose terrain to detect skidding or undesired yawing. Dec 6, 2021 at 21:00
• the mismatched speed of wheels happens in a regular car too. it's normally dealt with using a differentials. Dec 7, 2021 at 3:31

You have too many degrees of freedom to use it effectively without feedback. If you add a couple of potentiometers to detect the angles of the pivots, you should be able to compensate appropriately for it. Essentially the robot needs to know which way each wheel is pointing to determine which way and how much to try to turn it. The rest of "stable" steering (relative to the robot) will depend on whether or not you can process the feedback loop fast enough and your feedback algorithm.

In theory one person could try to learn how to process all 4 wheels at once while observing the robot for feedback, but to make it easy enough to call stable for the average person, it would involve creative controls. Humans tend to have 2 eyes and 2 hands. Steering wheel, gas pedal for car. Left, Right for tank drive. Are the twos a coincidence? I dare say I'm not cut out to be a puppeteer.

Yes, there are vehicles that have the steering by bending in the middle. Mostly graders and other heavy equipment though.

• They normally have hydraulics to "bend" the middle articulation. this would be passive. Also, the pivoting points are two in this case. Dec 6, 2021 at 17:23
• Well, my version has one pivot point, controlled by a single actuator which is less complicated compared to yours. Dec 6, 2021 at 17:24
• It surely is less complicated, an possibly it would work much better. It is just not what I have asked.. Dec 6, 2021 at 17:30
• @user1937747 then it may be worth considering that you are asking the wrong thing... Dec 6, 2021 at 17:32

You can use 3 gears in series - steering in the middle that rotates the gears on the axles in opposite directions by the force couple.

• This is actually a good idea, even though I was more interested in a non mechanical solution. Dec 7, 2021 at 16:56