You are currently viewing SemiWiki as a guest which gives you limited access to the site. To view blog comments and experience other SemiWiki features you must be a registered member. Registration is fast, simple, and absolutely free so please, join our community today!
I keep hearing that 3D sensors will be a break away sector over a medium-to-long horizon; assuming this is so (and I don't think this is an unwarranted assumption, if you agglomerate all the potential from automotive, industrial and consumer electronic apps), am I right to think that ams will be the key beneficiary (given their fairly comprehensive strategy of acquisitions in this area) - are there any well-placed competitors
What i had in mind were the sensors that Yole discussed in a report last year - which said (effectively 3D sensing and imaging in consumer applications is expected to increase the market size by a factor of 7 until 2022. So far as I can see the company best placed to take advantage of such growth is ams - though i may be confusing things
You have 3D sensors in phones these days which were barely there a year ago. That seems like one example of growth. Sensors in vehicles, factory equipment, other home gadgets (your TV, your bathroom or wardrobe, security systems, seem likely especially as costs fall and tech improves.
You mean 3D vision? From a software standpoint, I think most of it is opensource, although some companies are claiming proprietary I think those claims are pretty dubious. So if you are looking who might be competitively positioned from a semiconductor standpoint, I'd say companies that have IP in signal processing. Maybe Ceva would be a good bet.
No, the sensors in phones are not open source, not at all. Not the vehicle sensors either. There is a hotbed of innovation in range sensors and most of the IP is fresh enough it will probably be proprietary for years.
Even with simple stereo there are plenty of algorithms around, mostly not published. A lot of machine learning too, which is rarely shared directly (although the recipes for recalculating sometimes are, and you probably need to find your own data).
Thanks for the responses gents; I suppose I did mean 3d vision - ams refer to it as 3d sensing - 3D Sensing - AMS
Ultimately, what I am not sure of is whether ams' claims - that they have a significant headstart on IP in this field - is credible. And, if so, do we think that 3d sensing is a major growth area for semis? I am a little dubious but would obviously defer to the experts in this case!
The AMS list on page 25 is fairly conventional by now (for example Kinect 1 used structured light stereo and Kinect 2 uses indirect TOF, iPhoneX uses structured light, Intel puts a TOF sensor in PCs and laptops, …). Those are unlikely to be open IP yet, not as practical implementations. AMS may have been used as a manufacturer for some of those and so may have reciprocal rights reflecting its own contributions. They don't say anything about how their LIDAR works, just claim a high power VCSEL. But I am not aware of any VCSEL source running at 1.55 microns which is probably where you need to be if you want to pass safety regs on a 100W laser source, even if you are in a rotating device. Still, they quote 2021 as a commercial timeframe for that work, which is a sane and patient date.
Looks like a good set of products. Lots of competitors, of course.
