Toshiba Corporation announced advances in robotics that take the company closer to its concept of a "life support partner" -- a sophisticated robot able to deliver human-centric technologies that provide assistance and support the elderly and young children in the home and in such public places as shopping complexes
Truly effective robots have to be able to recognize multiple individual voices and multiple instructions from different directions at almost the same time. And if they are to be real partners, able to accompany people wherever they go, they have to be able to recognize and follow specific individuals. Toshiba's new robots draw on the company's image and voice recognition technologies to showcase these capabilities.
The robot that can distinguish and recognize voices from any direction, ApriAlpha _v3, also known as Apri sharp ear, integrates six microphones into its body to assure omnidirectional voice capture. Toshiba's proprietary voice signal processing technology allows the robot to recognize the direction of multiple speakers and what they have said.
The robot that can accompany people, ApriAttenda, first of all identifies an individual with its visual sensor and high-speed image processing system by instantaneously recognizing the registered color and texture of the person's clothes. The ability to follow a moving individual requires capabilities going beyond facial recognition. ApriAttenda has to detect and recognize the person and calculate his or her position from a complicated real-life landscape of fixed objects and moving people. Toshiba's new algorithm allows the robot to extract a particular individual from cluttered backgrounds, and to find and reconnect with the person if it loses visual contact.
When the captured subject moves forward, ApriAttenda moves forward. When the person stops, the robot moves to point beside the subject and also stops. ApriAttenda constantly calculates position and the distance to the subject and adjust its speed to maintain a certain distance from the subject. As the subject moves, an integrated ultrasonic sensor checks for obstacles in the robot's way, and the robot negotiates these while maintaining visual contact with the person. If ApriAttenda loses that visual contact, it calls to the person as it searches for and tries to restore the contact. The company targets commercialization of fully-fledged life support home robot integrating all these capabilities in five or 6 years time.
to Daily News