5.1 KiB
| title | tags | ||
|---|---|---|---|
| 02-intro-pervasive-and-ubiquitious |
|
Pervasive and Ubiquitious computing
Natural user interfaces
- applications off desktop
- physical linteraction moving away from mouse/keyboard/display
- natural acations used as input to systems
- better learnability, easure of use. ⇒ support tasks without changing them
- support special needs people
[!INFO] computer we wear communicate with more locally installed les ubiquitious computers
context awareness
- apps are aware of the environment around the user/the app and can adjus their behaviour to suit it
[!INFO] computer should behave in a more natural way [!INFO] also aware of our current goals
automatic capture and access
- automatically record and store things to remove the burden from humans and allows us to focus on things we are better at
Examples
[!INFO] phone already knows your location
[!INFO] uses electroculography: eye tracking without a camera. wanted to know if you could build a context aware systems based on eye movements: copy, read, write, video, browse, NULL. They achieved 70%-80% accuracy, WITHOUT using complicated machine learning.
[!INFO] measure cognitive load using pupil dilation. with high cognitive load they show less information on screen to stop from distracting you.
[!QUESTION] if they block other stuff while we have high workload, how do we decide which important things to show
[!INFO] camera is class: idea to show automactically that everything is caputred, audio recording are matched to slides, annotations are shared with the whole class, as well as individual comments. able to split up video to see which parts of the slides are being talked about when. collects a lot of data that can be used later
[!QUESTION] auto sharing annotations sounds very easy, i think we should make the default that annotations are shared
[!INFO] cameras constantly running, recorded his whole life. he is controversial, recording other people.
[!INFO] if everything is recorded there is an ethical dilemma, is contanst recording what we really want.
[!INFO] same thing: camera constantly running: but neck camera
[!INFO] toll gates scanning number plates: initally kept private. can we use it to solve a crime though?
Tangible computing
- Directly-manipulable physical interfaces to data and computation
- ‘Pure’ form of ubicomp in that there is no computer to be seen
[!INFO] message "captured in balls"
[!INFO] physical telepresence. physical interaction remotely
Calm computing
[!INFO] interactive surfaces and spaces. physicalise information. wire was moved according to network traffic
Input and Interaction
Google ATAP. track fingers using radar
[!QUESTION] Is it always tracking your finger? what if you do something accidentally
Wearable computing
Thad Starner, starting using weable computing such as glasses. more interested in showing information, than capturing it
Challenges
- What’s difficult about Pervasive and Ubiquitous Computing?
- Requires understanding of sensors and new sensing technologies
- Noisy inputs
- Sensor fusion
- Wireless communication channels
- Power consumption
- Context is only a proxy for human intent
- Unpredictable ussage context in particular for mobile devices
- Require novel user interfaces, limited/different I/O capabilities
- Lack of standardization in interface patterns
- Privacy & Security











