Hi all,
I am trying to make a excavator depth measurement system. Currently the market exists and it appears its all over the map on how its done. Some use laser and a stick in the ground to measure how deep the trench you are digging. Others say they use accelerometer/gyro's. The system I want to focus on is without lasers so I don't have to use outside devices or worry about sunlight. Issues I am running into so far is
GPS - Not Accurate to this scale Laser/TOF/Sonar/Ultrasonic - Can't use since you would need external devices mounted outside the excavator (pain to keep moving as you go along and sun)
Barometric pressure sensor - Subject to much to outside envolpe and not accurate enough Accel. Sensor/GYRO - Seems to be the only option that I have seen. However as everyone seems to find when measuring distance traveled it just isn't too accurate. Than I heard of dead reckoning which seems more inline however this seems to have issues also. Inclinometer sensor - Seems the most promising and being used in some systems
My only other thought is if I use two sensor that determine the angle of each part of the boom and I know the boom length can I determine the depth through this? I really don't understand the math too well but I envision a triangle since there is two parts to the boom and the actual excavator is the third which is constant?
Can I just do one sensor and get away with dead reckoning? The only library I could find on this was using a wheel encoder and not an accel. sensor
Update: Using accelerometer and this formula I found I was able to find the inclination. After seeing another system use this and he says in the video its done by inclinometer sensor. Stuck on the math part though on how to figure out the depth based on angle
In reality your perfect use case drawing is almost never going to happen. you are not sure the excavator will lay flat at altitude 0, as you dig, the excavator will move around and climb piles of stones or whatever you are excavating, sometimes being at weird angles …
Pros define a given level with a rotating laser and have a laser beeping sensor sliding on a graduated stick and will lock the sensor up on the stick so that the beeping occurs at the right depth. The pulsating laser/sensor is working even in the sun and at quite a distance (tens of meters).
You need also to have a worker to measure the depth in multiple places in the hole you dig. It’s a two man job.
Thanks for the reply. When using a system like this you can't move the excavator without re-calibrating each time so that takes care of what you are saying. Yes there is laser levels that can help with grade, and some systems use the laser level in their depth checking, but it comes with many downside such as needing to move it and in certain light conditions it will not work correctly. The depth checking system is just for simple trenches for laying pipe. it can be plus orminus .4 inches. I am well aware of using two men as we do it all day long, but it is great on long ditches to not have to have a guy check every 5 seconds and with grades changing in long runs it really helps.
The rotating laser is not attached on the excavator so you definitely can move around. The one I’ve seen being used (I did lots of construction work recently removing an old construction 40m x 20m wide digging 1m down) and their laser was working great in all lighting conditions. It’s an auto leveling box and at time they were relying on the signal 30/40m away without any issue.
I can see how this in more work if you need to dig a constant slope of 2% indeed.
But knowing how deep is the hole if your excavator is not staying flat and at a known level won’t help either. It’s like with voltage, You need a référence point…
As I mentioned already we do digging all day long. There are lots of issues with the laser setup, which is why if you look at all these systems the laser is typically only a second option. The laser does have to be moved from time to time. You have to remember also that this is being built on consumer grade parts to keep costs down so even if I wanted to go laser the length wouldn't be as good as these expensive systems. Knowing how deep your hole is does help because you zero it out when the bucket touches at the grade and each time you move the machine which sits on grade. Anyway not looking to debate the use class. I know it works in the field and was just looking to make my own for something I am doing. I actually think I found the solution after posting this. It comes down to a triangle. It gets very complicated but basically if you can figure out the angle and length of two sides which the lengths are static than through the math you can get the third length and angle. Than the sensors give you the angle of that missing side. If you know the length and angle it turns into a calculation of an arc radius since it's pivot point for that angle stay attached to the machine. I will post the update tomorrow.
I’m not debating your expertise there - I’ve just witnessed how the laser is being used when I did a massive construction project recently and it worked fine. May be they have an expensive system - seemed pretty good to me (auto leveling, all day battery, waterproof, long distance …)
I’m more debating the value of knowing how deep you go in absolute terms as your maths are set in the excavator coordinate system - so if you move it up and down you will reach different absolute depth whilst the math will show constant depth. (Not to mention how you drive the "end bucket")
So my question is more how do you compensate for the terrain not being flat / at zero level ?
You need to work out multiple triangles, one for each link in the chain from ground to tread, tread to pivot, pivot to boom, boom to arm, arm to bucket, and bucket to ground.
Each joint needs a correction compatible with its motion. It alos twists, so depending on the two-dimensional inclinations of the tread platform, the angle of the waist joint will affect the depth/Z of the bucket.
If you know the roll and pitch angles of the tread platform, the waist angle, the arm angle, the stick angle, and the bucket angle, along with the lengths of the links, calculating the x,y,z of a point on the bucket is just a pile of sines and cosines of those angles and links collected in one 4x4 transformation matrix:
The inclination of the digger and it’s height above the dig is an issue , anyhow ..
After you’ve sorted the maths …
If this is a real project…
You also have to think of the harsh use environment - hobby sensors won’t cut it and will quickly fail
Look for suitable sensors , how you would wire them and cost out your system , how it will be powered and results displayed . My fag packet suggests £5-10k project. ( £5-£800 per encoder , plus linkages )
Don’t forget you can change bucket attachments.
At one time I designed devices to look for buried gas pipes, cables and even waste pipes, which then lead to the dilemma of having found the offending buried object and roughly how far down it was, getting a digger to dig the trench to a specified minimum depth. Solution was man with big measuring stick , but it didn't stop me thinking about it.
My idea at the time was a boom arm from the top of the cab, with a laser sensor on the end, which pointed at the bucket.
It doesn't matter what angle the cab is at, as the calibration point is from the end of the boom arm, to the bucket resting on the ground any change in that distance, is the distance below ground level.
As I said, not implemented but an idea.
Inclinometers are probably the best option, given they measure with respect to gravity, but you would need sensors at each point of joint flex. You know the distances of each arm, so calculating how far down you have dug is the easier bit. Calibration is also just a matter of resting the bucket on the ground at the starting dig point for a zero.
The guy that just did dirt work at my house for the new shop had a laser unit that sat out on the ground on a tripod that kind of looked like surveying equipment and a sensor that just stuck by a magnet to the side of the arm. It was dirt simple. The laser put out a fan like a laser level and the magentic thing on the arm just picked it up. He did a measure with the bucket against the ground to calibrate and then after that all he had to do was touch with his bucket and it gave him a depth reading.
