The Active Home provides a natural domain to develop perceptual, cognitive, mobility, and manipulation assistance in the independent living context. It addresses barriers such as multidisciplinary collaboration, complex interactions between people and environment and need for dynamic human-system interfaces. The difference between the Active Home and previous instrumented and "smart" living spaces is that the focus of the Active Home is actuation: moving and doing actions for and with the inhabitants. In the Active Home, we will explore systems that can assist in cooking, eating, cleaning in the kitchen, home cleaning, dressing, personal care, using the bathroom, home maintenance, etc. We will also explore systems that assist a spouse or other caregiver manipulate a user, such as rolling over someone in bed, helping someone sit or stand, or transferring a person between a bed and a wheelchair.

The Active Home will include perception, reasoning, interaction, and action modules that greatly extend the capabilities of traditional home automation and security systems. Initial projects for the Active Home will include people- and object-tracking systems, and a mobile lightweight manipulation system to assist with object manipulation in getting objects, assisting with food preparation, and assisting with cleaning. This project is transformative in that it embeds home assistance that is intelligent, aware, reactive and proactive. The system will feature a high level of autonomous planning, as well as sophisticated perception and multiple levels of interface (ranging from high level verbal to low level joystick or touch screen control).

Our transformative intelligence goal for this engineered system is the combination of a high degree of awareness and decision making in a personal mobility and manipulation system. The user will be able to control the system using high level commands such as "get me a drink", "go to class", or "go to mom" and the system will be able to avoid obstacles or move them out of the way, to recognize and track people in a crowd, as well as avoiding moving people. Another transformative element is the level of manipulation assistance we will be able to provide to users who need it. The system will be able to get objects even if a cabinet door needs to be opened, for example. The system will be able to adjust clothing and feed a user. In order to improve time efficiency the system will prepare for the anticipated action, before the user provides a signal to act.

Our transformative capability goal is to build a device that can go as fast as powered wheelchairs with integrated limbs to manipulate objects, the environment, and the user. Design challenges include 1) making a device that can be both small to get through narrow passages and yet still have a large enough reach to be useful, supporting reach in a full hemisphere around the device; and 2) developing limbs that are fast enough to get objects and do other tasks quickly, soft enough to touch the user and attend to her, and strong enough to be used to move the device over obstacles and climb up stairs, move the device into and out of vehicles, and transfer the user in and out of the device.

The Safe Driving family of engineered systems is a cluster of projects focused on improving quality of life through technologies that support independent driving. Person and Society is integrated into all projects in this cluster. Like aging in place, the transition from driving to not driving can often occur due to external factors that have no relationship to the driver's actual ability. This is mostly manifested as an issue with payment - either for equipment or training. Therefore, technology alone is not sufficient. Likewise, it is also important to address payment, policy, and privacy issues.

Rather than take an incremental approach, the Safe Driving projects are looking to the future and are seeking to leverage trends in vehicle design, components, and population demographics. For example, vehicle sensing capability is increasing dramatically (e.g., GPS, adaptive cruise control, collision warning systems, etc) and the ability to package additional sensors is improving at a steady rate. However, these sensors are not ideal for certain measurements and adding new sensors is not free. Therefore, the team is leveraging decades of cutting edge work in robotics, active vehicle safety, and automated vehicles to measure key driver capability metrics with low-cost sensing. If successful, such systems will allow a driver rehabilitation specialist to examine long-term trends rather than rely on self-reporting and sparse observation.

Presently available cognitive aids are simplistic, providing only scheduled reminders and rote instructions. Future virtual coaches will actually monitor user performance of activities and provide appropriate feedback and encouragement. As the user's abilities change, the coach may reduce the number of verbal cues as the subject learns, or provide increased support as needed. A care provider could upload new capabilities to the virtual coach, as required, potentially without even an office visit. Virtual coaches also provide constant and consistent observation/monitoring, even on a real-time basis, thereby extending the clinician's guidance beyond episodic patient examinations.

Our virtual coaches will evolve starting with primary measurements, such as physical activity, location, biometric data and biological signs. Examples using the primary measures include balance and exercise coaches. Longer-term versions will add visual observation and local interpretation of physical motion sequences and behaviors in the performance of a routine task. Based upon observations, other capabilities can be derived such as user activity and intent. Examples for this stage would work in cognitive and social application domains. The coaches share basic science including sensing, logging, and modeling. Modeling will range from statistical to inferential. Machine learning will be utilized to adapt as an individual's situations and abilities change. Elements of game design will be utilized to keep the users engaged.

Virtual coaches reduce the barriers of current modes of service and support delivery and assuring safety. Our coaches will focus in particular on the loss of independence. Postponements of even six months in moving from independent to assisted living will have dramatic cost savings. Virtual coaches will progress from recognition of activities to identifying intent to autonomic behavior.