Frequently Asked Questions
What does spatial resolution, accuracy, and repeatability mean for a LIDAR sensor?
Like a camera, a LIDAR has XY spatial resolution, accuracy and repeatability. In a camera and a LIDAR, the spatial resolution is defined as the size (or field of view) of a pixel projected through the lens into space, typically defined as an angle (degrees) or solid angle (steradians) units, and many of these pixels are combined into an image. Higher spatial resolution means the pixels map to smaller features in space like cracks in a roadway or text on sign. For stringent sensing applications, both cameras and LIDAR also can define a spatial accuracy, which is how well we know the true direction that each pixel is projected into space. Spatial accuracy may be affected by lens quality, or in a scanning LIDAR, by the the ability of the system to measure where it is pointing at a given time. Spatial repeatability is less commonly discussed, but is the ability of a system to measure the direction the pixels are pointing again and again. In a scanning LIDAR, one driver of spatial accuracy and repeatability is the rotary encoder's ability to measure an angle consistently, combined with the system's ability to compensate for slight thermal shifts across temperature.
We’ll have a blog post on this topic coming up to go into more detail.
What does range resolution, accuracy, and repeatability mean for a LIDAR sensor?
Unlike a camera, a LIDAR also has range resolution, accuracy, and repeatability. Range resolution is defined as the smallest difference in distance that the LIDAR can measure and is driven by the system's rate of digitizing an optical signal as well as other internal signal processing. A LIDAR that digitizes an optical signal at 1 GHz has approximately 15cm of range resolution which is insufficient for many applications. LIDAR systems either use very fast electronics or employ signal processing techniques to improve their range resolution into the single digit centimeters. Range accuracy and repeatability map to the mean and standard deviation of a set of measurements taken for a target at a fixed range. LIDAR manufacturers generally provide worst case accuracy and repeatability specs derived by repeating these measurements for fixed targets at many different ranges. In a time of flight LIDAR system, the range accuracy and repeatability are decoupled from the target range and this is a significant advantage over continuous wave LIDAR systems or stereo camera systems where accuracy and repeatability degrade significantly with range.
What does range resolution have to do with the sensor's ability to discriminate targets?
Range resolution should not be confused with a system's ability to discriminate between two targets that are very close in range to one another when the same laser partially illuminates both in the same measurement. For instance when imaging a tree canopy and partially hitting a leaf and the ground below. The ability to discriminating two targets at different ranges with the same laser pulse is driven both by range resolution and the duration of the laser pulse in time. As a rule of thumb, two targets cannot be resolved if there is less than half the length of the laser a laser pulse between them.
Do your sensors have multiple returns?
The need for multiple returns originated from previous generations of LIDAR sensors having (1) low resolution, and (2) high beam divergence. Regarding resolution, our OS-1-64 product has four times the vertical resolution than the previous best typical sensors used for surveys on drones. Regarding beam divergence, our sensors have >2.5x lower beam divergence than comparable sensors. For example, our laser “spot size” at 100m has a diameter of 25cm, versus other sensors that have a 60cm spot size or greater, resulting in worse spatial accuracy, and the need for multiple returns as a stopgap to develop a point cloud with high spatial accuracy. We take the strongest return with more beams, and with more focused beams.
Is your sensor solid state? Are you working on a solid state version?
The OS-1 and OS-2 product families are mechanically scanning. Historically mechanically scanning LIDAR suffered from a number of issues - form factor (they were large/bulky), mass (the hundreds of discrete parts added significant mass), price (complex designs increased cost), quality (scanning with high mass and many discrete components created short product lifetimes), and availability (the sensors were difficult to make).
Ouster has a new approach to designing and manufacturing sensors that addresses these issues. Solid state sensors come with their own downsides including inherently smaller field of view and resolution. Our sensors are small enough to seamlessly integrate into a variety of locations on vehicles (e.g., fascia, headlight, windshield), drones, and other robots.
We don't comment on future products and features, but we’ve got an incredible roadmap ahead of us - we are serious when we say we are bringing Moore’s law to LIDAR.
What applications can Ouster LIDAR be used for?
Our LIDAR can be used for a variety of applications. Here are some examples of sectors and use cases of current customers:
- Drones - mobile mapping and surveying; detect and avoid
- Industrial automation - autonomous robots, factory automation
- Mining - mapping, automated vehicles and equipment
- Civil engineering - site surveys
- Architecture, engineering and construction - indoor and outdoor mapping
- Utilities - mapping of utility lines, vegetation management
- Automotive - autonomous vehicles, Advanced Driver Assistance Systems
- Cities - smart cities
- Universities - robotics and other research
- Robotics - various robotics applications
Can your sensors be used on a drone?
Yes, we are significantly lighter, smaller, and lower power consumption than other LIDAR sensors, and also enable higher performance - enabling many new applications on drones.
Can Ouster LIDAR be used for mining applications?
Yes, we do not have any restrictions on our customers buying our products for use in the mining sector - or any sector / use-case for that matter - and that will never change!
Do you include any software with your sensor?
We include ROS drivers, which are open source and posted on GitHub here.
How can we interface with the sensor?
The sensor can be queried and operating modes set over TCP and a GUI. Data output is over UDP over gigabit ethernet.
What are the requirements to integrate the OS-1 into my system?
It depends on the system, but generically two options exist to capture data:
- Have a wireless data uplink from the sensor (mounted on a robot or drone) to a computer to save or render the data
- Have a computer onboard the robot or drone with storage media to store data for later processing/use
What type of computer do I need?
If just storing data, a basic prototpying computer that costs $100-300 should work. If processing data real-time, you’ll need a more powerful computer. We’ll test different computers over the coming months and provide more guidance.
How can I time sync my sensor, for example with GPS?
Each measurement is accompanied by a timestamp, and we offer three methods for time syncing:
- Our sensor supports, and we recommend, IEEE 1588 Precision Time Protocol (PTP) for all customers, as it offers the best time accuracy and syncing in a system and especially multi-sensor architecture
- The sensor accepts a GPS pulse via a dedicated pin
- We have internal clock derived from an internal accuracy, low-drift oscillator
What is PTP and why do you support this protocol?
Where can I get more information about Ouster ROS driver?
Our github repository is here, and our drivers are open source.
How do we mount the sensor?
Mounting is 4x M3 screws. We also have locations for two high tolerance 3mm dowel pins similar to Velodyne. We include exterior mechanical drawings in our manual, and offer adapters that allow for easy attachment to existing brackets that use legacy LIDAR sensors,
Where does the name Ouster come from?
We’re a bunch of SciFi geeks and are big fans of the duology Hyperion and Fall of Hyperion by Dan Simmons. If you have read the books, you’ll realize why we chose to pay homage to the Ousters, but if you haven’t, we don’t want to reveal spoilers.
What is the background of the company?
The company was started by Angus Pacala and Mark Frichtl, who have been in the LIDAR for robotics industry for many years.
What sets Ouster apart from other LIDAR in the market?
On the technical front, five things: performance, price, quality, form factor and availability.On the company front, a customer-centric approach and high quality customer service. We pride ourselves on being responsive to customers, rapidly integrating features customers request and making sure our product satisfies their needs.