C haptEr 9 design Environments and systems
394
Consider a first example involving relatively simple cognition
levels. Biosensors could be placed within an environment to detect
biohazards that might result from a terrorist action. Within an
integrated electronic system, information from the sensors could
be analyzed and interpreted in relation to human tolerance levels,
and various alerts or warnings would be communicated, even direct
mitigation systems called into action. In terms of Figure 9.69, this
scenario would be illustrative of “single-behavior” sensory detection, coupled with a very simple expert system (a basic “if-then” formulation) and implemented through a direct mechatronic model
(discrete component and system control). A more complex example
is an “intelligent room” environment designed for individuals with
special medical needs or who are recovering from an operation.
These rooms might have a whole host of means for monitoring
personal vital signs, event recovery measures, general mental and
physical condition states (e.g., awareness, ability to function), indicators of whether medicines have been taken regularly, whether
the patient has had visits, whether the patient has moved around
and how (mobility assessments), and a whole host of information
concerning surrounding physical and environmental states, such as
temperature levels. Facial characterization and gesture recognition
systems might be needed for further assessments or as a way for
nonverbal patients to communicate.
The initial sensing goal in the above example is to detect and
measure everything needed for an externally located medical group
to access a patient’s state and progress. If fully automated, expert
or artificial intelligence systems would analyze and interpret information and, if necessary, cause specific actions to be evoked (e.g.,
at a low level this might be a voice communication to a patient to
“Take your 11:00 a.m. medicine,” or, at a higher level, could cause
a medical team to be dispatched). In this scenario, not only are
large numbers of sensors involved, but the interactions of multiple
behaviors must be analyzed and specific actions evoked. This same
room might make use of self-cleaning and antimicrobial surfaces
(see Section 10.2).
Though some basic ideas and aspirations are fairly straightforward,
not all these kinds of interactions and responses can be achieved at
the moment. Limitations exist throughout, including in sensory and
computational components that are more the focus of this book
than are related cognition algorithms, which remain extremely
troublesome. As we move from specific applications, such as the
medically oriented room we’ve described, and more into general
intelligent environments, there is less than clarity on what we really
394
Consider a first example involving relatively simple cognition
levels. Biosensors could be placed within an environment to detect
biohazards that might result from a terrorist action. Within an
integrated electronic system, information from the sensors could
be analyzed and interpreted in relation to human tolerance levels,
and various alerts or warnings would be communicated, even direct
mitigation systems called into action. In terms of Figure 9.69, this
scenario would be illustrative of “single-behavior” sensory detection, coupled with a very simple expert system (a basic “if-then” formulation) and implemented through a direct mechatronic model
(discrete component and system control). A more complex example
is an “intelligent room” environment designed for individuals with
special medical needs or who are recovering from an operation.
These rooms might have a whole host of means for monitoring
personal vital signs, event recovery measures, general mental and
physical condition states (e.g., awareness, ability to function), indicators of whether medicines have been taken regularly, whether
the patient has had visits, whether the patient has moved around
and how (mobility assessments), and a whole host of information
concerning surrounding physical and environmental states, such as
temperature levels. Facial characterization and gesture recognition
systems might be needed for further assessments or as a way for
nonverbal patients to communicate.
The initial sensing goal in the above example is to detect and
measure everything needed for an externally located medical group
to access a patient’s state and progress. If fully automated, expert
or artificial intelligence systems would analyze and interpret information and, if necessary, cause specific actions to be evoked (e.g.,
at a low level this might be a voice communication to a patient to
“Take your 11:00 a.m. medicine,” or, at a higher level, could cause
a medical team to be dispatched). In this scenario, not only are
large numbers of sensors involved, but the interactions of multiple
behaviors must be analyzed and specific actions evoked. This same
room might make use of self-cleaning and antimicrobial surfaces
(see Section 10.2).
Though some basic ideas and aspirations are fairly straightforward,
not all these kinds of interactions and responses can be achieved at
the moment. Limitations exist throughout, including in sensory and
computational components that are more the focus of this book
than are related cognition algorithms, which remain extremely
troublesome. As we move from specific applications, such as the
medically oriented room we’ve described, and more into general
intelligent environments, there is less than clarity on what we really