I would like to discuss the following configuration.
For a project i would like to mount 4 anchors on a driving trolley. A person is holding a tag by hand. In this case the tag can calculate his position xyz based on the fixed anchors on the trolley, calculate the distance and send this to a electrical wheel which drives the trolley to the tag based on distance and angle.
For this i have bought the MDEK1001 kit, mounted 4 anchors on a ‘trolley’ and made a handheld tag. I have configured the devices in a network using the app. And got successful readings.
My problem and question is how to improve the given position results. I have a interval of 200ms in the tag which sends his data through uart using lec command. This gives many NaN position results and thus bad practice.
In a normal configuration the tag is moving between the tags. in my config the tag is always outside the tags area. (hope this makes any sense). Is this the reason for bad pos calculation?
The 4 anchors are mounted on the trolley in a rectangle from 800 x 400mm and the tag is moving in a circle with a diameter from 0-4000 mm around the trolley.
If this bad result from the dwm_loc_get function is because of tag is moving outside of the anchor rectangle i would like to know if there is a different approach to do this?
Does this demonstrate the basic operation you want?
An array of anchors on a mobile platform can locate accurately a tag that is generally outside the lateral bounds of the anchor array. This mode of operation is often used when we build machine proximity systems, say forklift or heavy machinery protection systems for humans near by.
The system operates based on multiple tag to anchor range measurements. It works very well with 4 anchors, and can work with 3. When only 2 anchors hear the tag, it can still do a good job of locating the tag based on past history resolving the mirror position ambiguity.
The code shown here is Ciholas proprietary running on Ciholas hardware, so not immediately available for an MDEK kit. This does prove it can be done with a DW1000, though.
Mike Ciholas, President, Ciholas, Inc
3700 Bell Road, Newburgh, IN 47630 USA mikec@ciholas.com
+1 812 962 9408
Hi mciholas,
Great! Yes. this is exactly what i’am looking for. I assume you are measuring distance from each anchor to the tags and do the computing of the location in the PC, right?
I would like to test such a configuration with the modules which i have (MDEK Kit) and build a proof off concept. When succeeded it needs to be setup as commercial product with the right licenses etc… I assume your company can deliver the products for this?
Correct. The functionality can be embedded in the system without a PC which is what we do when we design this into a customer’s product.
The cost to have us port the code to the MDEK system greatly exceeds the cost of buying our DWETH and DWTAG devices. Thus the least expensive way to try our code shown in the video is by using our hardware.
That is our primary business, delivering designs to customers which embed UWB location technologies. The clients pay for licensing the technology from us and the custom engineering to make it fit their application. The client then can build their product and sell it.
Mike Ciholas, President, Ciholas, Inc
3700 Bell Road, Newburgh, IN 47630 USA mikec@ciholas.com
+1 812 962 9408
The video demonstrates our MultiRange algorithm, part of the Ciholas collection of UWB related technologies we can apply to designing custom UWB systems. The video was mostly aimed at applications where UWB is used for machine or vehicle proximity sensing, say to protect a human worker from getting too close to the machine.
The video uses our DWETH anchors which use a Decawave DW1000 chip.
Mike Ciholas, President, Ciholas, Inc
3700 Bell Road, Newburgh, IN 47630 USA mikec@ciholas.com
+1 812 962 9408
Hi behc0004, I did setup 4 anchors with 1 tag. The tag is calculating x,y,z, and send it out through the serial port. Somehow the tag has many bad results in calculating his position which results in NaN output. I cannot determine what the reason is for these bad results.
Meanwhile i have been playing with the DMW simple example program but havent got better results yet. I think it would be better to have each anchor measures his distance to the tag and send it out through spi or uart. With these distances i can start do some calculatings myself…
Perhaps there are some examples available for this approach?
It is impossible to convey the dozens of man years of UWB development we’ve done in a forum post.
Truly useful and practical UWB system development is a non trivial exercise, often far more complex than it appears at the start.
If I could find employees who would do the work for free, I’d be willing to share it for free. As that is not the case, I have to convert their work into revenue to pay them, so our code is not open source.
Mike Ciholas, President, Ciholas, Inc
3700 Bell Road, Newburgh, IN 47630 USA mikec@ciholas.com
+1 812 962 9408
Generally, we don’t use the Decawave modules and instead prefer to build our own chip down devices or our own modules. This allows us to optimize for antenna performance, size, features such as adding an LNA, and so forth.
Using a module is definitely easier, but we are focused on performance. If your requirements are satisfied by the Decawave module reception range, then that’s the way to go. As you exceed that, it gets progressively harder to engineer.
Mike Ciholas, President, Ciholas, Inc
3700 Bell Road, Newburgh, IN 47630 USA mikec@ciholas.com
+1 812 962 9408
