Alternative to Hokuyo 10LX

[vikas narang]

Hello guys 

I was looking to see if there are any alternatives to Hokuoy 10LX sensor . It is 1600 USD .  I am making this car as hobby . 1600 sounds little expensive . If you could recommend similar sensor whihc is at lower price will be helpful . If you could also share pros and cons will also help me to make decision ,

Thanks
Vikas 

[Jaroslav Klapálek]

Hello!

You can look for RPLidar A3 (for example). It is half the price, but also half the frequency. On the other hand, I think that 20Hz could be enough. Unfortunately, I have not tested this sensor. (Only A2s on Turtlebot, but that is not very helpful :) )

However, I can give you some clues for selecting another one:
 - frequency is very important; without frequent data from the environment, your SLAM will be very sad :( ; slowest lidar I have tested was SiCK TiM 561 with 15 Hz. It worked "somehow", but it was not definitely ideal.
 - field of view -- larger surrounding area gives your SLAM much more information to provide precise localization. >180° is really good, 360° is probably not that necessary.
 - maximum range -- Hokuyo is safely working up to 10m (but it can detect certain objects up to 30m) and for the competition (and basically indoor tracks) it is enough. However it depends on your testing environment.
 - angular resolution -- smaller steps ~ more detected objects; but that is basically it. If your target is not to detect poles/legs from chairs/... you don't need that fine laser scan. I believe that everything is working with even (close to) 1° resolution.
Last thing to think about is connection method to your car. In case you will stumple upon USB3 (if that exists) lidar, think a lot about that. There is only one USB3 port on the Orbitty board, and the main target for that should be a camera. Lidars can be connected other ways, e.g. via ethernet as Hokuyo. (It is a little bit pain to set up properly, but then it works flawlessly.)

I know that this wasn't the thing you have asked for, but I hope that it will help you.

J.

P.S. Hokuyo was already selected as a quite good, "cheap" lidar. Therefore, looking for a less expensive device, you will have to sacrifice some of the parameters, for sure.

Dne neděle 2. srpna 2020 1:44:03 UTC+2 vikas narang napsal(a):

[dfar...@gmail.com]

I am in a similar situation. I am doing this project as a personal learning experience. My thinking is that I am going to think of my racecar as 'f2tenth' Similar to size and function as F1tenth... but on a budget with an accompanying performance decrease. 

[vikas narang]

Thanks Jaroslav .  Now I reached to a point of ordering LIDAR sensor. I was just looking at the specs for this sensor and to my suprise I found that this LIDAR sensor is only an indoor sensor and will not /doesnot work outdoor . I also read some reviews 

from people and they also had issues using this outdoor . Do you see this as a concern ? or F1tenth was originally designed to operate indoor only ?

Thanks

Vikas 

[Jaroslav Klapálek]

I would say that F1Tenth was always designed as an indoor competition. On the other hand I have tested the car during various lightning conditions (worst case would be a direct sunlight through glass roof during morning summer) and I had zero issues with Hokuyo.

As far as I remember, we have also tested the car outside in a back alley and it was still working (detecting nearby walls). I think that outdoor won't be that big issue, especially if you build the track using some borders.

Maybe others could also share their experience with this. :)

J.

Dne pondělí 26. října 2020 3:25:52 UTC+1 vikas narang napsal(a):