At CEATEC Japan 2005, Hitachi displayed a voice-operated prototype Mascot Robot (nicknamed Bot-chan), an intelligent user interface for AV equipment. Standard remotes are not necessarily the best interface for the growing number of television channels, their programs, and the ability to record and watch content on demand (TiVo). Remotes have become increasingly complex as a result, only complicating things further.
Hitachi developed the Mascot Robot for quick and easy access to content through speech recognition, speech synthesis, image recognition, and text analysis technology. The robot analyzes the user’s viewing history to effectively alert you when a show is on, record your favorites, and recommend shows you might enjoy. Simply tell the robot (which can hear you from up to 2 meters away) what you’re looking for and it can find it for you.
Besides its cute voice, the animated eyes on its LCD display and the robot’s “rabbit ears” can move to help express the robot’s “emotions”.
Essentially, Hitachi was trying to develop an interace that would reduce a user’s stress through intuitive operation using only natural language. The robot on display relied on an external PC but a stand-alone model was being developed, with an expected consumer product by 2008. The technology has three main features:
(1) Analysis of user preferences through viewing history
Automatically learns user preferences based on viewing history, and without the need to enter keywords, you can tell it to find, play, or record specific programs. Information is retrieved from the EPG (Electric Program Guide) which stores the program title, genre, cast, episode summary, length, frequency, as well as whether you have watched the show or episode before.
(2) Adaptive personal service
Using facial recognition technology, the AV Mascot Robot knows each family member, and adapts its recommendations to best serve each individual. For example, it sees you sitting down to watch television, recognizes who you are, and suggests the show you always watch on that day and time.
(3) Multimodal recognition
A user’s needs are transmitted through speech recognition, speech synthesis and integrated multimodal facial image recognition. Through voice-recognition and synthesis, the Mascot Robot delivers an interactive natural language interface. In addition, the robot estimates the user’s location using each microphone in the microphone array and uses facial recognition technology to look at the user and make eye contact to assist in friendly interaction.
Hitachi developed at least 2 different prototypes of Bot-chan – spot the differences in the photos. JVC had displayed a robotic AV mascot, called the J4, at the previous CEATEC show. Toshiba has also pursued the idea of a robotic universal remote with its ApriAlpha and ApriPoco robots. All of these were developed prior to tech like the Wii remote, which is cheaper and could be successfully applied to standard AV equipment.
Impress Robot Watch | CNET Japan