Hi, currently i have a Loomo device and it has an Intel RealSense ZR300. So i would like to ask how to get the object distance in java (android development)?

[MartyG]

Although Loomo has a RealSense ZR300 camera in it, it uses its own Android SDK by Segway. Loomo has its own support website.

https://support.loomo.com/hc/en-us

[WLIM0]

​Is it possible to use its SDK to detect object distance from intel realsense sensor to avoid obstacle?

[MartyG]

According to the Loomo documentation, you can use Loomo's mobile app to set Loomo to detect obstacles and avoid them when it is in Follow mode.

https://support.loomo.com/hc/en-us/articles/360003503972-The-Follow-Function-in-the-Mobile-APP

[WLIM0]

​Ya, we have using the ultrasonic to detect it but the ultrasonic have some limitation which is its vision is only 40 degree and will cause the wheel collision on some obstacle. Furthermore, for the loomo sdk on follow mode, it only can detect person. If not a human, it will return 0.

[MartyG]

Loomo has a Docs page but that requires a Loomo account login so I cannot tell if there is anything there that could help you program your own obstacle avoidance routine. It does have an 'obstacle avoidance API', implying that this function is custom-programmable. I would recommend logging in to the docs and seeing if you can search the information on Loomo's direct-control 'Avatar' mode says that obstacle avoidance is not currently supported in that mode though, so it may not be accessible outside of Follow mode in the API either.

Looking at the hardware spec of Loomo, its processor (Intel Atom Z8750, also known as 'Cherry Trail' architecture) means that it might be usable with RealSense SDKs if you wiped the storage and did a clean install of RealSense SDK instead, using it as a RealSense robot. That would mean losing all the special features that Loomo supports though, including the mobile app control, so it may not be worth it.

A less extreme approach that would preserve the Loomo software would be a dual-boot setup where you boot the RealSense SDK from a USB storage stick placed in the USB drive.

[WLIM0]

​Yes, we have login to the website and go through the documentation, but still cant get the thing for detect obstacle via Intel RealSense Technology. May I know which part is the obstacle avoidance API? Sorry I cant find it

[MartyG]

I do not have a login for the Loomo website unfortunately, and are having problems with accessing support pages that are publicly available.

I managed to track down the GitHub site of the tools that Loomo provide to interface the robot with ROS robotic control software, but nothing specific on the obstacle avoidance functions.

https://github.com/segway-robotics

[WLIM0]

​Oh, thanks buddy. Is it you have develop this robot before?

[MartyG]

There are some products that contain Intel RealSense cameras but are produced by other manufacturers and have their own SDK software that is not developed by Intel. Although Intel does not officially provide support for them, I do try to provide assistance with these special SDKs if there is documentation or other guides to research, but ultimately it is difficult because there is no way to directly do tests without having the product. I do truly understand the frustration in these cases.

[WLIM0]

​So is it means that Loomo already redevelop the SDK for intel realsense in this robot and currently there is no way to do what I want?

[MartyG]

As Loomo said in their docs, the live-control Avatar mode apparently does not currently have obstacle detection. I guess the logic is that if the robot is being live-controlled then obstacle avoidance is not needed because it is within the ability of the user to stop the robot if they see something in its path. The most recent code update to their GitHub was 6 months ago, so I do not know whether their SDK is still receiving updates.

Given that Loomo seems to have strong support for ROS integration, your best option may be to find a ROS obstacle avoidance module. There are plenty of resources on this if you google for 'ros obstacle avoidance'.

[WLIM0]

​Ok, thank you.