detecting viruses, germs, pathogens early and perhaps in real time with wearable devices will be very useful and popular, also having better, more sensitive environmental monitoring would permit moire accurate weather modelling and more efficient use of home environmental controls like AC, Heating etc.
Sensors provide computers (and their owners) information about the world around them. Computers won't reach the next level of intelligence without a comprehensive set of sensors. Ultimately the biggest impact of sensors may be in intelligent machines, robots and eventually androids. However in the end it is all about helping humans and the other branch will be enormous - directly adding extra sensors to humans. We already have a huge advantage over computers with our many built-in sensors (many more than 5 by the way) but adding sensors for things we can't natively sense or sensors that improve on our built-in senses will increase our abilities and make life safer and more productive. This has already started with increasing numbers of sensors in cell phones and the trend will continue with more person-borne sensors and more sophisticated person-borne sensors. In the short term our machines and toys, such as cars, houses, appliances, sports equipment and cell phones will start sprouting more sensors, but eventually our personal sensor space will extend far beyond our bodies and it won't even be clear what is person-borne and what is just extracted from external sensor networks. To answer the original question, simply ask yourself where you spend the most money on things - that is where the most sensors will appear in your life. So houses, cars and cell phones.
I would have to say medical, with the ability for your doctor to check your condition if you call in. Or if you had a heart attack or stroke your sensor implants could call the ambulance.
The next big use for sensors will most likely be for automated driving, that will be the first attempt at robots doing more human tasks. If that experience fails it will be years before you see more adoption of robots.
One of the best sensors might be to stop relying on technology so much, and use our own senses more.
Humans seems to be relying more and more on technology to tell them information, that in some instances makes little or no difference.
We seem to be obsessed with multitudes of facts and figures, and then wonder why we missed the clues that our bodies tell us.
I've seen localised weather conditions change, and catch out people who used technology to predict the day, rather than some common sense, and preparedness.
Basically without the technology they fall into some unknowing mode and can't think for themselves.
Stick them where there is no electronic coverage, and they are unable to function.
IMO automated driving is a myth.
It is not suitable for retrofitting into every vehicle on the road and doesn't actually resolve anything.
It is fine in controlled environments, but it relies on high cost or high maintenance sensors, making cars very expensive and even un-affordable.
The next issue you have ... Is each car autonomous or do they rely on other vehicles having similar or compatible units.
How do you check (and prove) they are working correctly, and the cost of replacement when one is detected faulty or worn out.
(If you think they will be cheap, try pricing a replacement ECU for your car, or some of the ABS or throttle sensors)
To prove how easy just one part is to fail have a look at this
Now tell me you are happy to be on the road with an autonomous vehicle !.
So what kind of applications is a good question.
I do know the device size and raw component cost will reduce, but whether that translates into cheaper consumer items is a different question.
Some of the technology is already available, but the R&D and IP associated with it, makes it an expensive 'extra' or a selling feature, rather than standard equipment.
The 'Forget Me Not Challenge' showed there are some very effective solutions and the Open Source market has a big part to play in the future of sensors.
If anyone doubts Open Source look at the success of Arduino.
While some would design it differently, it has a large following in many areas.
