Hi, I'm Chris Campbell, Lead Solutions Architect at ROVR. AMA!

@Bitbeard The LC (LightCone) will be available for pre-sale on the ROVR official website starting June 16th, with an estimated delivery in September.

@GeoCKC Any plans to sell the LC through authorized resllers, like Helium Deploy?

@Xrave Yes, definitely. For the second batch of LC devices going on pre-sale on June 16th, we're increasing the power input voltage range to directly support both 12V and 24V. We've also optimized the device's weight and mounting bracket, reducing the weight without compromising safety. Plus, we've improved the cables

@Xinran_ROVR any learnings about the data quality itself? Anything to improve about the lidar system?

@Xrave That's a great question. Deploying the first batch of LC devices really highlighted situations we hadn't anticipated. For instance, some users placed the LC inside their vehicle, behind the windshield, and our data quality detection module hadn't accounted for this scenario before.

The LiDAR system itself still has some issues, particularly with low-temperature heating. The LiDAR we currently use automatically turns on self-heating below 0 degrees Celsius, drawing up to 45 watts. This brings the total LC power consumption to 62 watts. We're actively discussing this with the LiDAR vendor to see if there's a solution.

@GeoCKC yea it’s not really feasible to put it behind the windshield unless you calibrate out the windshield, and even then windshields don’t have coatings to optimize for IR transmission = not really worth it to spend time to improve…the user’s rewards being reduced will provide the push for them to externally mount.

Hard to know what part of the lidar system is impacted by the low temperatures. Lasers generally have more efficiency at lower temperatures and the feedback systems should adjust power to stabilize the output. Depending on the photodiode material temperature can have different effects, similar with the laser, for example some of your operational range drops below the lasing threshold.

My guess is the main issue is with the MEMS scanning mechanism and nonlinearities in the servo control loop. MEMS are fairly temperature sensitive and because this is a scanning lidar system the control loops might be able to compensate properly, resulting in reduced range or frequency response.

@Xrave I believe the value of TX is long-term.

TX is lower in cost and easier to install, which means it will generate significantly more data and with greater freshness than LC

@sabakas Not at the moment. We've solved the multi-view transformation issue using a generative deep learning model, and we'll be releasing some sample videos soon.

@bob_carsyn more than 700km/week

@GeoCKC If you are hitting that , but it’s the same route for 60% of the miles is that still effective for the project?

@bob_carsyn Road conditions can change, there might be road construction, accidents, or traffic jams.

Currently, traveling the same route with a one-week interval is still meaningful for the project, and rewards will be calculated separately for each instance.

@bob_carsyn freshness gets turned on July 1

Each road will see a 50% reward reduction for every 2 times additional collections.

Road Visits Reward Ratio

1st - 2nd 100%

3rd - 4th 50%

5th - 6th 25%

7th - 8th 12.5%

…… ……

• Road Revisits are calculated globally.

• Road Revisits Records Reset Weekly: Revisit records will reset on the Monday of each week at 00:00 GMT.

• $ROVR will be transferred within 1 week after data upload, data older than 2 weeks won't be accepted.

• This mechanism will take effect on July 1st, 2025 at 00:00 GMT

@Xrave As always, you are a wealth of knowledge Xrave, a gentleman and a scholar!

@GeoCKC Is this in terms of an LC or TX driver. It sounds like the project only prefers highway drivers with numbers like that, but the data is surely important for cities as well right?

@DePIN_Pete I was wondering about the same thing

@DePIN_Pete We've prepared sufficient inventory for this drop, and each unit comes with a phone mount. If you place your order through HeliumDeploy, they'll ship directly from the available stock, so the delivery time will simply depend on the shipping — no need to worry.

@djsupafly It depends on the level of shaking. During video preprocessing, we apply software-based stabilization across consecutive frames

@GeoCKC 20～50 pixels

@djsupafly for the TX any camera shaking/blur only really affects the sharpness quality metric, which is only 20% of the overall rewards. It can technically also affect contrast although likely not enough to change your rewards. Brightness, dynamic range, and noise level are immune from motion blur effects, although if we want to get super nerdy on camera image histogram analysis you could actually argue that some motion blur actually helps your rewards for these metric because the blurring (averaging) of these metrics pushes the final result into the higher reward gradings for each of them.

And obviously RTK accuracy, which is the largest contributor to your rewards, is unaffected by the camera performance.

@joinsamp yes we have coverage for Canada, the US and Europe

@DePIN_Pete Not at the moment. This is because we need to transmit some data in real time for analysis and preprocessing, which requires a cellular network.

It’s technically feasible to create an LTE version of the device, but it would require users to pay an additional monthly data plan, which we believe is not ideal for scaling.

@bob_carsyn https://carnation-caution-df7.notion.site/ROVR-data-quality-69b874170d85429580bc5c78ccfeb7e3

@bob_carsyn This is the detailed dimension for data quality evaluation. Please Check~ ROVR Data Quality

@bob_carsyn section 5 https://rovr-network.gitbook.io/rovr-docs

They changed to a simpler grading scale but in the backend they use the following metrics, based on camera quality: sharpness, noise, brightness, contrast, dynamic range, RTK coverage, and for the LC-only the additional metrics of point cloud density and point cloud outlier distance

@william On the device, we run ROS2. Sensor data (from the camera, LiDAR, IMU, GPS, etc.) is organized into ROS2Bag format and stored on an external hard drive.

@DepinMasterMind Yes it is

@GeoCKC Ok so how is it planning to deploy as many units and match HiveMapper’s coverage?

@DepinMasterMind We’re gradually expanding our marketing channels, although we recognize this is still an area where we have room to grow. That said, we’re committed to increasing the project’s visibility and improving service quality.

Very soon, we’ll also be launching our Fleet Management feature, which will allow more users to actively participate in the ROVR project.

On the data product side, we’re actively working to make breakthroughs and expect some good news soon. I believe this will be a positive development worth sharing.

@Xinran_ROVR Also I think one advantage of ROVR is that it still has a lot of tokens to distribute versus a project like HiveMapper that has been on the market for a longer time!