Suspension for Mobile Robotics

[Jones1987]

Hi,

What methods are used for applying suspension to robotics? I'm designing a surface surveyor, so I want minimal disturbance created by the surface is it travelling.
I thought tracks would remove some disturbance rather than 4 single wheels. However I believe these will still receive some disturbances.

However for rough terrain, is there anyway I can apply some form of suspension to which the robot will be able to travel and collect data without body pitch causing incorrect readings?

 

[FlexGunship]

Hi,

What methods are used for applying suspension to robotics? I'm designing a surface surveyor, so I want minimal disturbance created by the surface is it travelling.
I thought tracks would remove some disturbance rather than 4 single wheels. However I believe these will still receive some disturbances.

However for rough terrain, is there anyway I can apply some form of suspension to which the robot will be able to travel and collect data without body pitch causing incorrect readings?

You'll be limited, essentially, to an active justifying system using accelerometers to determine angle of travel. There are no passive methods which can guarantee upright travel without entirely separating the body from the chassis. If this is an option, then simply suspend the portion you want stabilized in a freely rotating ball-joint with a counterweight at the bottom. It won't protect against sudden jarring movements, but you can ensure it stays upright that way.

 

[Danger]

One thing that I must point out that the nature of the gravity field in which you want this to operate is extremely important. You can get away with a lot lighter components and spring rates at 1/4g than you can on Earth.

 

[AlephZero]

If you go for the idea in #2, you will also need something to stop the "pendulum" swinging about. For example it would swing even on a perfectly flat surface when the vehicle changed speed. You could use a gyro to stabilise it, or some sort of mechanical damper. You might need to stop the vehicle to let the platform stabilize, each time you take a measurement.

The opposite approach of making fairly rigid platform with some way to measure its orientation seems easier.

To avoid the long term drift problems with accelerometers, you could use load cells or strain gauges to measure the force components of the weight of an something, in directions relative to the vehicle orientation. For example build a pin-jointed "tripod" shaped like 3 edges of a regular tetrahedron on your measurement platform, put a mass on top, and measure the forces in the 3 legs. The resultant of the 3 forces will always be downwards relative to the ground.

 

[Danger]

Do you have any idea of how irritating it is to someone like me when you make sense? I thought not.
That was a great post, Aleph.

 

[FlexGunship]

If you go for the idea in #2, you will also need something to stop the "pendulum" swinging about. For example it would swing even on a perfectly flat surface when the vehicle changed speed. You could use a gyro to stabilise it, or some sort of mechanical damper. You might need to stop the vehicle to let the platform stabilize, each time you take a measurement.

You don't need the pendulum to swing free. Imagine it attached to a short section of steel re bar.

 

[AlephZero]

You don't need the pendulum to swing free. Imagine it attached to a short section of steel re bar.

We may be thinking about different ideas, in the absence of any pictures, but...

Your pendulum must be able to move, otherwise it's not a pendulum at all. If its size is of the order of 100mm to 1m, it's natural vibration period will be of the order of 1 Hz. So whatever makes it move, it will oscillate at that frequency till the motion gets damped out. I would think the force input from the vehicle on rough ground, or even accelerating decelerating and turning on a smooth surface, would have a frequency spectrum with a significant amount of energy at the same 1Hz frequency.

The "mass on a tripod" idea will also have its own dynamic response. The forces in the legs will oscillate, even if the structure is too stiff to see any obvious vibration, but it would be easy to make its lowest vibration mode say 100 Hz, which I assume would be high enough not to mess up what it is suppposed to be measuring.