Posts made by Alex Kushleyev

@groupo , thank you for the feedback and offer to help. Yes, please send us an email with some more details and we can discuss if there is an opportunity to collaborate.

Alex

@Roy-Azriel , the description of Starling 2 Max states that it uses VOXL2, as well as lists other major components and options for camera configurations : https://www.modalai.com/products/starling-2-max?variant=48216084087088

Please let us know what Lidar you are considering adding.

Alex

@aheyne , please send us a message via contact form. We don't sell this part separately as a product, but will help you with a replacement.

You will need the M0177 flex cable (as the print states on the cable).

https://www.modalai.com/pages/contact-us

Alex

@Piyush-Singh , can you please provide more details

• how exactly all cameras are connected to VOXL2 mini (including interposers, flexes), etc. A diagram would help a lot
• which kernel variant are you running on your voxl2 mini (2.0.0 or 2.0.2).

Are all the other cameras working if you disable the TOF sensor in the camera server config?

Alex

@Shantanu-Gupta please check documentation and video for VOXL2:

• https://docs.modalai.com/voxl2-voxl2-mini-system-image/#force-voxl-2-into-fastboot
• https://docs.modalai.com/voxl2-connectors/#buttons
• https://www.youtube.com/watch?v=IM3PnW7cipQ

Alex

@fhaltmayer , which wifi usb adapter are you using?

Hello @RoyAzriel ,

If you are going to be experimenting with adding different sensors (Lidar), then Starling 2 Max is definitely going to offer more flexibility due to the larger size and longer flight time.

You should consider the dimensions of the tunnel and make sure that the drone will easily fit. I have no idea how small the cross section of the tunnel is, but if it is very small (just several times larger than the diameter of the drone), then it would be very difficult to fly due to air circulating in a small space.

What Lidar are you considering adding and what is its weight and communication interface?

The illumination will depend on your specific application, depending on how far you need to see. You should be able to add standard LED lights. if you need to control the LED intensity, you would need to get an LED with built in adjustable LED driver and optionally control it using PWM output feature on the ESC.

Are you going to develop your own mapping and navigation stack for this application?

Alex

we just posted an update on this topic, please see https://forum.modalai.com/topic/5116/gnss-emi-mitigation-guidelines

we just posted an update on this topic, please see https://forum.modalai.com/topic/5116/gnss-emi-mitigation-guidelines .

@colombe-ai , please see https://forum.modalai.com/topic/5116/gnss-emi-mitigation-guidelines

@groupo and ALL, please see https://forum.modalai.com/topic/5116/gnss-emi-mitigation-guidelines

Dear ModalAI Community,

We are fully aware that many of our customers have been affected by very poor GNSS performance on the Starling 2 and Starling 2 Max vehicles. We really appreciate everyone's patience and trying to work through the issues and we apologize that it took so long to provide a formal analysis of the issue.

Our team has posted an initial document which captures the results out testing of GNSS performance on Starling 2 and Starling 2 Max vehicles.

Due to the number of components which have at least some effect on GNSS performance, the reported performance results varied across the vehicles due to slight changes in the configuration, such as which components were present / enabled, how many cameras were installed and what resolution / fps they were running and so on.

The actual test results for individual components and plots and videos will be added soon, as the work is still ongoing. Please sign up to receive updates from this thread, if interested in staying up-to-date.

Please note that most of the cameras on these vehicles had quite severe effect on the GNSS signal. In some cases, the solution is as simple as disabling the camera (like TOF, which mostly not usable outdoors, but was till running and causing interference). One big item remains is the IMX412 camera(s), which is core component of outdoor video streaming, but causes significant interference (up to 15dB loss!). We will be working on testing the camera settings to reduce this interference (as soon as possible).

The Starling 2 and Starling 2 Max vehicles are no longer advertised as GNSS-capable, although they may still ship with GNSS receivers installed (you can request it, if needed). A new GNSS receiver mast has been designed for the Starling 2 Max V3 (late fall of 2025) and is available in CAD (easy to 3D print).

Nevertheless, we believe that it is possible to make substantial improvements in GNSS performance without significant mechanical changes.

We hope the details provided in the document, as well as the ways of testing different components, will allow our customers to tailor modifications to their specific application.

We will be considering these findings for our future designs, however the current Starling 2 and Starling 2 Max vehicle design will remain mostly unchanged in the near future.

Those who are interested in improving GNSS performance in their existing vehicles are strongly encouraged to read through the document and implement the suggestions. Feel free to ask specific questions about anything related to this topic in this thread.

As the document is updated, we will share any major findings in this thread as well.

https://docs.modalai.com/emi-mitigation-for-gnss/

• ModalAI Team

@Windmill , if you use 1920x1080 resolution for your IMX664 camera, depending on other settings in voxl-camera-server.conf file, the 1920x1080 will be a cropped version of the full resolution 2704x1540. That would match the result that you got : 150 * 1920 / 2704 = 106.5 degrees of diagonal FOV.

Please share the camera config section for this camera and I can tell you how to fix it.

Alex

@sjt277 ,

The simplest approach is to use a separate USB-to-serial adapter for each ESC.

You can set up udev rules to map and distinguish the specific usb-to-serial adapters (using their unique serial numbers) to aliases in /dev/, such as /dev/esc0, /dev/esc1, etc.

Then run your testing using different USB devices to access each ESC (i am assuming you are using voxl-esc tools?).

For example, here is some relevant information: https://askubuntu.com/questions/49910/how-to-distinguish-between-identical-usb-to-serial-adapters

Does that help?

Alex

Hi @Matt69 ,

Yes, the shown ip address is a self-assigned address which happens if there is no ip address assigned by a DHCP server.

What wifi module are you using?

Does output of iwconfig wlan0 show that you are connected to your Access Point?

Alex

Hello @Tanner-Metzmeier,

We do not have a heat sink recommendation for VOXL2, however you can try off-the-shelf components, just be careful not apply excessive mechanical stress to the QRV5165 SIP while mounting the heat spreader to it (or removing it later).

Is there anything in particular that is causing VOXL2 to overheat (what type of processing)? We could try to help reduce the system load, if you provide more details on what you are actually running.

Alex

@Adnaan-Yunus , Thank you for reporting this. We will try to replicate the issue.

Do you have any more specific information that could help us? For example

• does rebooting VOXL2 allow you to connect from the same wifi client again (perhaps one time until it is cut off again)
• is there any way that you found to reset the wifi card to allow a new (previously seen) connection?
• have you tried setting the card to station mode and then back to softap mode?
• while the first client can no longer connect, can another client connect to the softap? I am wondering if the wifi card is still responsive or it does not accept any clients?

Thanks!

Alex

Hi @Jon-Brookshire ,

Please see the following post (just updated) : https://forum.modalai.com/topic/5052/starling-2-max-cad

Alex

Hello @Myles-Levine and All,

You have been very patient.. Sorry for the delay.

The Step file for the exact configuration you requested has been uploaded to our Assets page:

https://developer.modalai.com/asset/6

It should look something like (well, exactly like ) the rendering below. Please let me know if you have any issues with the model. It may take a few minutes to load..

Alex

Hi @jameskuesel ,

When you enable en_raw_preview, which is what we do when MISP is enabled, the preview_width and preview_height basically force the selection of a specific camera mode with those dimensions. This means that the camera will send a bayer image of that size.

After the image is received by the ISP (which does nothing else in RAW-only mode), then MISP consumes the raw bayer image and performs debayering on the GPU. Then misp_width and misp_height will be used to specify the dimensions of the output image, sampled from the original bayer image.

If the output dimensions have exactly the same ratio as the input dimensions, then MISP will perform (arbitrary) down-scaling only (no crop = same horizontal and vertical FOV). However, if the output dims have different aspect ratio, the output image will be a cropped (and downscaled) version of the original image, such that the width and height ratio is maintained (features are not stretched). The crop is selected to maximize the fit of the output image within the input image.

With that in mind, if you select 4040x3040 as preview resolution and select misp resolution such that both width and height are scaled by the same factor (does not have to be integer), you can calibrate the intrinsics using the small image and after the calibration you can upscale the intrinsics:

• multiply the principal point offsets by the same factor
• multiply the focal length by the same factor
• keep the same fisheye distortion coefficients (since they are a function of angle, not pixels)

See another version of this explanation in a different context : https://forum.modalai.com/topic/4900/running-qvio-on-a-hires-camera/12 (look for "intrinsics")

• Input to the EIS algorithm is the full bayer image that is defined by the preview_width and preview_height.
• the output image size is defined by misp_width and misp_height, just like when EIS is off. The misp resolution can be arbitrary
• an important parameter is misp_zoom, which defines the relationship between the original image and the output image, in terms of zoom level. if you set zoom = 1.0 with EIS enabled, then the misp output will produce the largest undistorted image that fits within the original bayer image
• there will be no deformation (stretching). the EIS ROI size will fit within the original bayer image and the zoom level is really controlling how much the FOV is reduced and how much stabilization margin you have.
• you can enable the ROI display (PIP or side-by-side) to see how the EIS ROI fits within the original full image
• the actual resolution of the misp output has nothing to do with how the EIS ROI fits within the original bayer image : the zoom value and fisheye undistortion dictate that. The output misp resolution just defines how many pixels the output image will contain, not the field of view (FOV).
• for output resolution, you can pick whatever you need for your application.

By the way, you can have multiple EIS outputs on the same camera concurrently (MISP supports up to 4 channels). So you can have one stream that is small resolution (for streaming over wifi) and another one higher resolution (and can be different aspect ratio) for high quality recording. https://docs.modalai.com/camera-video/electronic-image-stabilization/#using-eis-with-misp-channels

Please note that currently, when you use multiple misp channels from the same camera, the image is processed separately, meaning if you have a large resolution and small resolution misp stream, the small resolution is NOT generated by downsampling the output of the large stream (which could be a nice optimization, in some cases).

Please let me know if you have any other questions.

Alex