When we started Sense a few years ago, Chris, Ryan, and I had this notion that if we knew\u00a0in detail\u00a0what all of the electronics, appliances, and other electricity-using devices in a home\u00a0were doing, then\u00a0we could help people make their homes work better, by helping them to better understand energy\u00a0efficiency, know when things are broken, and get awareness of what is going on in their home from day\u00a0to day. We knew from our own experience that there were not really good solutions to this \u2013 we had\u00a0each tried running around the house with an electric meter to figure out how much energy every device\u00a0used, or calling the HVAC repair guy over and over again because the heat pump really didn\u2019t seem to be working quite right.<\/p>\n
Based on our past work in speech recognition, we thought the electrical signals may have enough\u00a0information to allow us to learn what various electrical devices in the home are doing just by carefully\u00a0monitoring the overall power into the house and then doing signal processing and machine learning to\u00a0figure out usage signatures for each device. If we could do that, we could then feed this information\u00a0into an application to give people detailed visibility into what all the devices in their homes were doing.<\/p>\n
So, we started digging into this, and sure enough, people had been doing research on this for a while\u00a0(starting with some work at MIT back in the 1990\u2019s). The concept was called \u201cload disaggregation\u201d and it\u00a0seemed like it shouldn\u2019t be too hard. So, we started building some early prototypes and collecting data\u00a0to see if the stuff we read about in the research papers would really work.<\/p>\n
Based on what we saw, we quickly realized that the technical challenge of making this work was much\u00a0more difficult than the academic research had made it out to be!<\/p>\n
The problems we ran into were actually somewhat familiar \u2013 when we started in the speech recognition\u00a0industry we had also started with technology from MIT (I was a researcher there before starting\u00a0SpeechWorks in 1994) and the difficulty we had implementing the technology was pretty much the\u00a0same. We had systems back then that worked \u201cin the lab\u201d \u2013 so would work well for a small number of\u00a0people speaking in quiet environments if they said the right things. But, the real world challenge was\u00a0MUCH more difficult. In speech recognition, the challenge is the huge variety of accents, speaking\u00a0styles, noisy environments etc.<\/p>\n
For load disaggregation, the problem is similar \u2014 the research shows how to make load disaggregation\u00a0work for small numbers of predictable devices in predictable environments. But in the real world, there\u00a0is a huge variety of devices (and therefore electrical signatures) and a huge variety of electrical noise in a\u00a0typical home. And, the signals we are able to measure is the combined effect of all the different devices\u00a0in the home \u2013 kind of like trying to do speech recognition with 30 people all talking at the same time.<\/p>\n
So, we\u2019ve spent the past few years working hard on making load disaggregation work for real houses \u2013using a lot of what we learned from the speech recognition world, as well as a ton of new techniques\u00a0that we developed in the process. The basic answer for how to handle all the variability we see (and this\u00a0was the same answer for speech recognition as well) is to use more and more complex models, add in\u00a0lots of data from the real world (so, lots of examples of the things we want to recognize), and drive it all\u00a0with machine learning algorithms to figure out the right models based on the data.<\/p>\n
We\u2019ve been making a lot of progress and now have gotten to the point where we are able to make\u00a0things work well enough for the initial products, which we started shipping at the end of 2015. The\u00a0good news is now that more Sense monitors are out there, we have an increasing amount of data\u00a0coming in, which is what allows us to make the technology work better and better \u2013 the positive\u00a0snowball effect has begun!<\/p>\n
We have a great team of people working on the modeling and machine learning techniques and are\u00a0excited about the progress we can make on increasing the coverage and accuracy of our models. And,\u00a0for those of you who are early customers, don\u2019t worry \u2013 our hardware\/software architecture allow us to\u00a0constantly update models in the field, so your Sense monitors will benefit from all of the progress we\u00a0are making on the technology front.<\/p>\n","protected":false},"excerpt":{"rendered":"
Our CEO Mike writes about Sense’s early days: what inspired Sense, the initial experimentation, what made Sense such a challenge to develop, and how the team’s backgrounds in speech recognition helped overcome those challenges.<\/p>\n","protected":false},"featured_media":1095,"template":"","meta":{"_acf_changed":false,"om_disable_all_campaigns":false,"_relevanssi_hide_post":"","_relevanssi_hide_content":"","_relevanssi_pin_for_all":"","_relevanssi_pin_keywords":"","_relevanssi_unpin_keywords":"","_relevanssi_related_keywords":"","_relevanssi_related_include_ids":"","_relevanssi_related_exclude_ids":"","_relevanssi_related_no_append":"","_relevanssi_related_not_related":"","_relevanssi_related_posts":"1","_relevanssi_noindex_reason":"","footnotes":""},"consumer-blog-category":[23],"content-type":[13],"featured":[],"class_list":["post-1093","consumer-blog","type-consumer-blog","status-publish","has-post-thumbnail","hentry","consumer-blog-category-sense-monitor","content-type-article"],"acf":[],"yoast_head":"\n