Context-aware systems are already part of our everyday life; like any technology that enables ubiquitous computing, their ease of use means that they essentially “disappear” or are no longer noticeable. Examples include an iPad switching the orientation of the screen when you turn it on its side or a car’s Electronic Stability Program and anti-lock braking system. Your satellite navigation is aware not only of your position, but of other contextual parameters such as light conditions, weather, traffic along your intended route, and time of day.
Automatic security lights that turn on when they sense the combination of darkness and motion are an example of a very simple, context-aware system. Next-generation devices with context-awareness in the home might include a smart valve that shuts off your water supply to reduce damage from a broken pipe as part of a smart plumbing system, smart AC outlets that learn device usage patterns and turn off unused appliances, or a sprinkler system in the garden that turns itself on in the evening when it senses that the day has been hot and dry.
Independence From the Main Brain
There’s a recurring trope in science fiction films including Independence Day, Ender’s Game, and Starship Troopers where the destruction of a mothership causes all its drones or subsidiaries to fail. Once the connection with the hive intelligence is severed, the machines either stop fighting or crash to the ground.
Currently, most IoT devices don’t have much in the way of inherent functional intelligence. The sensor observes variables, reports to a gateway (usually a remote cloud environment), and awaits a response. The gateway analyzes the information, combines it with data insights received from other sensors, makes a decision, and issues instructions back to the devices.
Smart Home IoT’s marketing manager Johan Pederson writes that this lack of independence, combined with the devices’ inability to “understand” data via machine learning, means that devices cannot make direct actions without calls to the gateway. This situation creates:
- Unwanted latencies
- Data security risks
- Lower device intelligence
- Low immediacy and flexibility
Pederson explains that the evolution – driven by advances in power, range, security, and interoperability – means that the next generation of IoT devices will buck a major technological trend by taking control back from the cloud and putting it in the devices themselves. He writes that this change will “move the paradigm toward a fully ambient environment, where adaptive intelligence is expected, invisible, and pervasive.”
Context Awareness in Industrial Manufacturing Applications
While most of the literature around context-aware IoT is centered on smart homes, an important paper written in 2018 by researchers Patrick Rosenberger and Detlef Gerhard explored context awareness specific to industrial applications.
Rosenberger and Gerhard proposed a framework with a number of industry-specific context types, including that a device should be aware of:
- Personal information: Who is the user? What is the users’ name, age, language, or habitual activity? What is their role in the company? This information can be used to provide sophisticated personalization: for example, a condition monitoring system could notify the exact maintenance worker required (rather than sending out a general error message) based on the detected error, the workers’ current location, and skills.
- Personal condition: Sensors can vastly improve safety by being aware of a users’ condition such as their heart rate or temperature. An example of this is a fatigue-sensing hat designed to battle trucker fatigue.
- Location: Context-aware sensors can recognize where their users are and respond appropriately. For example, an engineer using a tablet on the manufacturing floor will be presented with a certain set of options in the user interface (UI). However, if that engineer walks into the office instead, the UI will automatically change to a list of the tasks commonly performed in that second environment.
- Date and time: Being able to understand periods such as day and night, weekends, or special events in a calendar will help automate industrial applications.
- Environmental conditions: An awareness of temperature, lighting, humidity, and other environmental conditions means a smart device will be able to respond by performing an action such as sending a message to an IoT-enabled cooling device.
- Resource and resource condition: Being able to sense and understand the inventory and workspace; the “machines, tools and products used in production, assembly, and maintenance.” Sensors should also be able to classify the condition of resources (working or faulty) and production states (in progress or finished).
Making end-users’ jobs easier through context-aware environments means the designers’ job will be harder. Designers now have to anticipate and program a vastly increased number of situations and contexts in which their systems will be used. This may, in turn, drive up costs of IIoT devices, but the gains in ease of usage may well be worth any increase.