Today LightWare Optoelectronics released the SF30/B obstacle detection LRF that updates at an astonishing 36633 readings per second and has a range of 50m in all lighting conditions. The introductory price of the SF30/B is just US$350.
The SF30/B is the latest light weight laser product from LightWare designed specifically for use in unmanned vehicles. It includes both a serial port and analog outputs along with a micro USB port for configuration. The update rate, resolution and baud rates can all be set to suit most types of aerial and ground based applications.
The SF30/B is designed to operate either stand-alone as a conventional LRF or it can form part of a scanning LIDAR system to create two dimensional or three dimensional maps. The very fast update rate of the SF30/B makes it ideal for locating potential hazards along the flight path of a UAV and the 50m range gives the flight controller plenty of time to take evasive action.
Additional products in the SF30 family include the SF30/C with a range of 100m for use in larger UAVs, helicopters and light aircraft. LightWare also manufactures the SF10 range of precision AGL altimeters and the popular SF02 LRF. For more information visit lightware.co.za
thanks for your reply!
i hope i can review your product asap!
Hi sijongmu, we're doing a production run of the SF30's right now and will make a short video of the data collection and mapping capability once we're done.
i have been interested in your good product continously. but i can't find the demo film of SF/30B & 30C.
could you show me the film of performance in details?
@LD! Thank you very much! Very good and detailed answer! Now just learn fast and accurate navigation in the veri big room and do not see any similar devices from other manufacturers. Especially in this price category. Very cool! If my code comes to implementation in hardware, I will definitely buy your product. Already have ideas for scanning around the robot as it is done in lidars using a rotating mirror at an angle of 45 degrees.
Yes, we considered going to CAN but I'm worried about its performance with high speed LIDAR. The limit for CAN 2.0B is 1Mb/s and the SF30 is already using 50% of the transfer time without any of the CAN packet structure. Of course you can always drop the data rate and accept the loss of data points, but with multiple devices on the same bus or next generation products with even higher data rates in the pipeline, we keep coming around to the need for a solution that is data rate agnostic.
Perhaps the issue boils down to the difference between obstacle detection and mapping. Obstacle detection needs to be done at the highest possible speed but ideally should have low overhead for the host controller. Mapping, on the other hand, can be a little slower and analysis can be done asynchronously, perhaps even after all the data has been collected.
We see that there will always be a need for a controller of some sort to drive the scanning mechanism. This motion controller needs to collect additional data pertaining to the aiming direction of the laser and then combine this with the range data. We think that this same controller could take care of the high speed, obstacle detection process by sending a notice to the host only once something hazardous has been detected, but otherwise remaining relatively quiet. A mapping mode would also be available wherein the motion controller packages the full data scan for the host and makes it available via UAVCAN.
Whilst I understand the simplicity of a device connected directly to the CAN bus, our experience with LIDAR (we've made 7 generations of scanning lasers over 20 years) is that separating the interfacing and motor controller functions from the laser itself results in the most reliable system, especially once you factor in the effects of slip-rings.
This looks like a great device! I suspect we'll end up with one of these as part of our next Outback Challenge vehicle (mapping out the landing zone).
One concern I have is the UART interface. We keep running out of UARTs as more sensors want to use them. Have you thought about adding UAVCAN as an interface option? That would also allow for multiple of these Lidars to be used on one vehicle without running out of UARTs.
A UAVCAN <-> UART expander board is another option (eg. providing 4 UARTs via UAVCAN), but it sure would be nice if new sensors had UAVCAN support natively.
See http://uavcan.org/ for details on UAVCAN if you are not familiar with it.
@Evgeniy, to prevent interference between units looking at exactly the same spot the standard method is to "encode" the laser pulses by applying a mathematical function to the timing between the pulses. This is similar to spread spectrum (frequency hopping) radio transmission but instead of changing the frequency it changes the timing.
We're concerned about offering this kind of technology on a commercial basis because it also affects the detectability of the laser pulses, making them harder to locate by outside sensors. This would put the technology firmly into the military category and we don't want there to be any chance of this technology being abused.
The SF30 (as with most other laser products) will work just fine in multiple unit configurations as long as the lasers are not aimed in exactly the same direction. We recommend a 5 degree separation between laser beams so there is no chance of beam overlap, even if the mounting points flex due to normal operating loads.
Secondary reflections don't really happen with natural targets, the attenuation of the signal is just too high. Multiple signals from two targets with small angular separation can be handled in hardware on a first or last basis - the SF30 always gives the distance to the first return signal because this is clearly the safer option. Products like the SF02 will let you download a detailed image of all the signals over the measuring range.
Hi LD! Very cool device! Say please what if two devices will shine at one point (or in a area that is seen both devices)? If this problem is solved, how many devices can operate simultaneously?
I looked at that Beam divergence :: 0.2° and it 17cm on 50m. Perhaps it will not be a big problem. But still interesting nuances of the reflected rays when hit for example. Or for example, when part of the beam reflected from the object at a distance of 5 meters and a part of the object at a distance of 20 meters?
I've always believed the scanner can be attached to an aerial drone. If your drone can lift a gimbaled camera system with three motors, then the scanner should be NBD. My son, the programmer, is currently working on connecting your laser to ROS through an Uno. We'll keep you all posted.
Hey LD - great piece of kit. My start-up is working to enable this kind of system over the next 3 months for aerial vehicles.