This is an open source design (reference model) - it is free to use, and you can start today to build a device that supports these features. This device aims to help people with any of the following impairments: limited cognitive ability, dementia, limited hand (finger) coordination, poor eyesight, blind, or (partially) deaf. It is envisaged that such a device will be very useful to many elderly people, and it could benefit young children too. For those people such a device will enhance their quality of life on many levels (memory, performance, ability, health, safety, security, accessibility, etc.).
Draft for comments.
Please use the contact link at the bottom of this page.
This specifies the mandatory minimum requirements, and extra desirable features. All of the mandatory requirements can be readily achieved using current technology. It is envisaged that all of the desirable features will be viable (in a low cost device) within the next decade [2020s].
The device must support all of the following (to be awarded Cognitive Coprocessor status):
» Small enough to fit into a trouser pocket (recommended size: 12 x 8 cm, and a depth less than or equal to 1 cm).
» Robust: able to withstand a drop from 2 metres.
» Waterproof to a depth of at least 1 metre.
» Readable in outdoor sunlight by people with relatively poor eyesight: minimum text height of 0.5 cm; and high contrast colours for text and background (able to pass accessibility tests).
» Records and plays audio.
» Records and plays video. The device has forward and backward facing cameras.
» The device can store audio / photo / video notes and reminders. These are recorded with the in-built microphone and cameras. Note: no keyboard is specified for this device [because of the challenges associated with the range of user impairments (supported in a small low cost device)] - however see the desirable features below.
» Notes and reminders can be viewed (or listened to), created and deleted. No edit option is specified [see above challenges].
» Provides audio and video conversational functionality with a remote user (like a mobile phone, but using the simple user friendly interface defined here).
» Wireless connectivity (e.g. WiFi, 4G, 5G).
» Very easy to use: it must support elderly people with any of the above impairments, and young children. To cater for those with limited cognitive ability (and the blind) this means keeping the number of options available at any time limited to a maximum of three. Three dedicated physical buttons (labelled 1, 2 and 3 in digits and braille) are provided for this purpose. The options will be shown on the screen (which is not a touch screen); and read out audibly (if the user leaves this option enabled). A physical dedicated button labelled Menu will show the user the top level menu. If the device is off, pressing Menu will turn the device on immediately (an initial audio visual response must be presented within one second). It will then deliver a user authentication prompt. The device is turned off by an option presented on a menu.
» Alternative interfaces: Other solutions are acceptable provided they work well with all potential users, including those that are blind. For example, instead of physical buttons, it might be possible to have a touch screen that provides buttons suitable for the blind, by giving feedback as to which button their finger is over (e.g. audio feedback).
» User authentication cannot use conventional passwords. Given that this device supports people with dementia (and other cognitive limitations) authentication cannot rely on the memory of the user.
» The device provides a list of contacts, which can be quickly accessed. This allows a user to quickly contact people in the event of an emergency or panic attack. [For example, in the design it is shown on the top level menu as option 1.]
» The device must always provide a real-time response to every user action. In cases where some intensive (remote) processing is required the device should immediately indicate that it is processing the request. If the processing takes a long time then periodic assurances should be given to the user (e.g. every 10 seconds). (Ideally processing times should be less than 3 seconds, and never exceed 30 seconds.)
» The device, and its associated services, must be secure, fool-proof, and trustworthy - given that potentially vulnerable people will be using these devices. In practical terms this means that best practices (and beyond) should be deployed. It probably means that third party add-ons and "apps" should not be supported: only the native functionality provided and guaranteed by the manufacturer (including their suppliers/partners). It probably means preventing remote (online) software installation (and perhaps upgrades) - unless a secure, fool-proof, method can be guaranteed, given that vulnerable people will be using these devices. [A new innovative approach might be required here.]
The following features are desirable, but not mandatory. It is expected that all of these will be achievable, in a low cost device, within a few years.
» The physical keys available light up (i.e. options on the current menu). For example, a sub-menu with just two options would light up the 1 and 2 keys; along with the Menu key.
» An artificial intelligence (AI) agent. This will greatly enhance the functionality of the CC device and provide a better service to the user.
The AI could be incorporated into the user authentication process to intelligently use visual and fingerprint recognition (e.g. if the user had previously fell and injured themselves and so had a bandage obscuring part of their face, or if a user cut their finger (fingerprint) or hid it under a plaster, then an intelligent device might suggest an alternate method (face / iris / different finger)).
The AI could scan audio/visual notes and reminders to extract text and meaning. It could inform the user of a relevant note, depending on the context. It could notify a user when a reminder is due. It could tell the user the names of people they meet (memory aid).
A voice driven interface could be implemented. A thought driven interface could be implemented (with EEG headset). A silent speech (think the words) interface could be implemented with electrical sensors in a collar placed around the throat.
Intelligent context sensitive help could be provided for many scenarios. This would greatly enhance a user's quality of life on many levels (memory, performance, ability, health, safety, security, accessibility, etc.).
For some AI inspiration you might like this realistic sci-fi story:
Education 2049: Looking forward to the great journey.
Provided that all of the mandatory requirements are met, the manufacturer is free to design the device how they like.
Here are some recommendations that might enhance the overall quality and usefulness of the device:
» Adopt a common standard across all manufacturers
» Adopt the Good Product Design concept
The above schematic shows a device that contains a camera (along with one on the other side), a screen, Menu key, option keys (1, 2 and 3) and a fingerprint scanner.
The top level menu is shown with three options: contacts, all other functionality, and power off.
Other design approaches might be feasible, such as some of those relevant to virtual reality devices. A blind person might use a silent speech throat sensor and ear piece. Some might use smart glasses (with eye tracking) and an ear piece. Whatever the design approach it must meet the mandatory requirements.
This is an open source design (reference model) - you can freely use it now. Please provide a link back to this page.
If you have any questions, comments or suggestions then please use the contact link below.
Free advice is available to anyone that intends to develop a CC prototype.
Tue 6 Apr 19:20:44 BST 2021