"We have designed an Internet-of-Things (IoT) framework named WearSense that leverages the sensing capabilities of modern smartwatches to detect stereotypic behaviors in children with autism."
So said the paper by Amir Mohammad Amiri and colleagues  (open-access available here) and, I have to say, something that really piqued my [research] attention. Describing how authors managed to design and construct a smartwatch with the ability to "detect three behaviors, including hand flapping, painting, and sibbing [hitting themselves on the top of their head] that are commonly observed in children with autism" they report some preliminary findings.
When I say these are preliminary findings, I do indeed mean preliminary, as a two-phase preliminary trial included data from "12 healthy subjects aged between 23–33" and "two subjects (ages 15 and 16) diagnosed with autism." Aside from the implication that young adults with autism are somehow 'not healthy' (I think the correct terminology should be 'not diagnosed with autism/autism spectrum disorder'), you can perhaps see that much of the data for this study came from artificial, induced behaviours not necessarily produced by those on the spectrum - "The tasks that the subjects were invited to do included three different types for 20 s." I do have some other quibbles about the write-up of this study as per very generalised sentences like: "These stereotypic behaviors happen when a child is trying to regulate the sensory input from their surrounding environment."
But I don't want to take anything away from the potential of this kind of research and where, with a bit more study and refinement, it could take many areas of autism research and practice. Accepting the argument that stereotypic behaviours that can accompany autism are not always something that needs to be tinkered with, I can perhaps see a use for this technology when it comes to screening and assessment. If for example, this kind of technology could be applied to something like an ADOS assessment, you could perhaps see how there may be additional information to be garnered (and indeed, built up coincidental to the 'objectivity' linked to such an exam). Coupled with other technology in relation to things like gaze monitoring for example, the potential gets even more exciting. And then also are the potentials of this kind of tracking software in relation to monitoring physical activity and autism (see here for example) or even in the context of epilepsy occurring alongside autism (see here for another WearSense use). There may be lots more to see when it comes to such technology and autism...
 Amiri AM. et al. WearSense: Detecting Autism Stereotypic Behaviors through Smartwatches. Healthcare (Basel). 2017 Feb 28;5(1).
Amiri AM, Peltier N, Goldberg C, Sun Y, Nathan A, Hiremath SV, & Mankodiya K (2017). WearSense: Detecting Autism Stereotypic Behaviors through Smartwatches. Healthcare (Basel, Switzerland), 5 (1) PMID: 28264474