Popular Mechanics (South Africa)

The smarter home

How the Internet of Things will raise community IQ

-

Late on a Saturday evening your phone buzzes. You’re out at dinner and the conversati­on is going well, so you ignore it. Then another buzz. After the third one, you check to see what jokes are circulatin­g the family Whatsapp group and mute it. But it isn’t your family. It’s your fridge. You need to buy milk and cheese. Further down the notificati­ons list, you can tell that the TV was unplugged and, importantl­y, the airconditi­oning is trying to regulate the temperatur­e because the window is open.

You get home to realise that you were robbed. Completely cleaned out, even the fridge. Your house tried to warn you, but you didn’t listen. It’s okay, though, because the virtual neighbourh­ood watch can piece together the incident from the suburb’s security cameras. From this you can see the perpetrato­r’s car circle the block. Through deep machine learning the cameras have tagged feeds from earlier in the day that show how the gang staked out the neighbourh­ood, posing as bin scratchers.

Facial recognitio­n, number plate recognitio­n, pattern recognitio­n… all for free and open source to plug straight into an existing set-up. “Bring your own device,” says Ryan Beech of Ryonic Robotics and now Seventh Sense IOT. He designed and built the MEIO and is ready to take his open source platform to market. “You can bring a Raspberry Pi or Arduino, download the source code and be ready to go.”

Beech describes himself as a serial technology entreprene­ur. His greatest passion is to put South Africa on the technology map by creating world-class solutions. “IOT immediatel­y caught my attention when I came across it in 2013,” he explains. “I spoke to Bokkie Fourie and Shera Eshmade about IOT projects and applicatio­ns because of their specialist background­s. That is when Seventh Sense IOT was born. The rest, as they say, is history.”

The Internet of things is a concept in which everything is connected to a local network or a wide area data connection. Everything is not necessaril­y connected to the Internet. Autonomous cars, for instance, will communicat­e with each other in real time to make traffic more efficient; you’ll be able to control your TV via your smartphone when your au pair wants to change to a different streaming service for your kids; and, importantl­y, your house will be able to self-diagnose maintenanc­e and security issues. “The MEIO app and website will be a one-stop portal to view and control all IOT devices. It will turn data into valuable informatio­n offering tangible insights,” says Beech of his latest project. “This will eliminate the need for 15 different apps to gain access to all your IOT Devices.”

It’s an ambitious plan to give users control over all their connected devices. Not even tech behemoths Apple, Google and Samsung have managed to cast a net that wide. But that’s because they’re trying to sell you a product and lock you into an ecosystem. MEIO is free to use, open source and even comes with a dedicated support team who are the actual developers. The idea is to let people decide which devices to connect and

what functional­ity those devices will be endowed with.

It’s this democratic approach to connected objects that makes MEIO so intriguing. It transforms IOT into a platform that interprets data in a useful way instead of performing a manufactur­er-ordained, pre-planned, limited script of functions. The level of artificial intelligen­ce is limited only by your coding skill and imaginatio­n. It’s actually intelligen­t AI.

We last saw Ryan Beech helming the Parkhurst off-grid project, which has subsequent­ly morphed into the Parkhurst Smart Grid project. Naturally, MEIO is the common thread as the platform on which all the projects will be built.

The first phase is smart grid management. When enough households have converted to renewables, the suburb can establish a microgrid and control energy sharing and distributi­on between residents. After that, the 2 000 households and all businesses in the area will pool informatio­n resources and convert into a smart town, where visitors to the area can easily locate parking or get traffic updates, waste management and recycling receives a smart upgrade, and the town produces a digital neighbourh­ood guide.

The smart security is the most interestin­g, though.

“This [virtual neighbourh­ood watch] is a big part of our Parkhurst project and we are aiming to allow residents to patrol the streets from the comfort of their own homes. This will be done via MEIO as well,” says Beech. “Every resident will allocate two hours a month to watch the feeds and ‘patrol’ the streets. A panic button in the app and on the website will allow them to send a proactive response vehicle when they see something suspicious.”

Though there are other off-site monitoring options with some AI baked in, Beech believes that the intelligen­t scene recognitio­n combined with a central control room solutions are already outdated. “The latestgene­ration cameras all come with onboard storage that eliminates the need for a DVR, and they also have on-board video analytics to send events only to security companies or police. This eliminates the need for a big control room as footage is sent only if an event if flagged,” he explains. “MEIO also gives residents or business owners access to their video feeds and can, in similar manner receive footage only when preprogram­med events take place.”

Beech has built the deep learning AI himself with the express purpose of recog- nising patterns. Like the aforementi­oned route the burglars took to reach your house. The idea is to organise the security systems better than the criminals are organised and turn the war on crime into one that favours preparedne­ss.

The obvious red flags are when criminals infiltrate the system, posing as residents. As with all solutions that favour a sharing and honesty economy, it seems vulnerable to attack from the inside. Another problem is server space. For MEIO to work optimally on a large scale, the platform needs room to store all the data while it feeds the algorithms. What Beech has considered very well, though, is how to scale for the numerous local businesses in the area. “Our initial core focus is on industrial and commercial clients to obtain data from various types of machines and equipment. But [at the same time] we are developing the platform for the consumer and hobby market to enable makers and garage tinkerers to build IOT devices using Arduino and Raspberry Pis that they can then link to MEIO. We are aiming to develop an online code base where people can download embedded code for their devices to ensure seamless connectivi­ty to MEIO. MEIO portal will also have preconfigu­red devices for sale to connect anything from your pool pump to your fridge.”

Seventh Sense IOT is already developing hardware specific to the platform, most recently branching into manufactur­ing low-cost electronic boards for a number of new markets. These new markets probably

included some of Ryonic Robotic’s existing client base who are already using MEIO to varying degrees.

“All our robots are Iot-ready and our customers will be able to access the data they acquire using our robots through MEIO,” says Beech. “We [Ryonic] are also developing Africa’s first locally designed and developed industrial automation robots that include six-axis robot arms and cobots [collaborat­ive robots that work in unison with a human workforce and require more stringent regulation]. We will use MEIO as a remote access platform to view equipment status and vitals in plants and factories remotely all over the world.”

It would be easy for Beech and Seventh Sense to become as user-hostile as its competitor­s, given the extensive automation Ryonic is capable of. He does assure us that this is not his intention. “Most companies force you to use only their hardware and software, which is a frustratio­n we want to alleviate. We are also aiming to developed low-cost solutions as imports are expensive due to the South African exchange rates.”

These user-oriented cost-saving measures include a printable QR Code system to trigger IOT functions working alongside capacity for Nfc-enabled devices. This allows for cross-platform flexibilit­y and leaves room for growth if a third party hardware manufactur­er can afford to equip each device with an NFC circuit.

“My focus to date has been on developing world-class robots and IOT platforms,” he tells us. “We are, however, planning to approach select makers and makerspace­s once we are complete with MEIO’S beta testing in early 2017.”

February 2017 is pencilled in as the launch date for MEIO. The platform is well positioned to make the deadline, but Beech is conservati­ve in his thinking. His biggest hurdle at the moment is the high cost of data storage in local data centres. Seventh Sense is in discussion­s with service providers to find a way around this obstacle. There wouldn’t be much sense in bringing this product to market when the user experience is hampered by frustratin­g glitches, brought on by inadequate servers.

On top of the considerab­le effort that has gone into bringing the platform to a point where it can compete with the world’s best and brightest, Beech’s vision for the impact of this project is quite noble. “The real benefits of the Internet of things is in reduced energy consumptio­n, less food and water wastage, better securi- ty and ultimately a generally more comfortabl­e and improved living environmen­t. I want to put Africa on the internatio­nal technology map, deliver a world-class product locally and at an affordable price.”

When you think about it, a self-diagnosing home will save you money in the long run by intelligen­tly reporting faults before they happen. The self-aware smart home could also protect your valuables with an array of intruder detection methods and a robust incident escalation structure.

The counter argument to empowering the things around us is that it leaves us open to malicious attack. Or worse, HAL will refuse to open the pod bay doors.

 ??  ??
 ??  ??
 ??  ?? In its infancy, the deep machine learning component of MEIO’S virtual neighbourh­ood watch is behaving exactly as expected – seen here keeping a watchful eye over Beech’s garage. The live testing in Parkhurst has helped in rapid prototypin­g.
In its infancy, the deep machine learning component of MEIO’S virtual neighbourh­ood watch is behaving exactly as expected – seen here keeping a watchful eye over Beech’s garage. The live testing in Parkhurst has helped in rapid prototypin­g.
 ??  ??
 ??  ??
 ??  ??
 ??  ??

Newspapers in English

Newspapers from South Africa