Evolving autonomous home

Inventor:  Streczyn, Alex., San Francisco, CA 94118
Applicant: NA
Issued/Filing date: March 24, 2000
Application #:  US2000777143909
Ipc Class:  G06E 001/00;
Current:  382/155; 706/013; 706/019; 706/025; 706/027; 706/062
Field of Search: 382/155; 706/013; 706/019; 706/025; 706/027; 706/062; 244/159; 244/161


A home system optimized as an adaptive entity that learns from both verbal and nonverbal input.  It is an continuously learning entity drawing information from residents' verbal input as well as other sources such as input from various appliances and regulators throughout the home and the internet.  Instead of an increasing database, the home entity will evolve behavior that utilizes an artificial life program that adjusts an evolving  matrix network whose fitness parameters are defined by the resident.  Based on the strongest positive feedback from the resident, and eliminating those that recieve weaker feedback, better adapted behavior will evolve.  It draws on previous experience, user commands,  and on the internet in real time based on previous learning processes and commands.


  1. The "training" process is in control of the resident so that the home entity can be customized to the resident's needs.  The initial training will configure the operating system of the home entity.
  2. As the resident's needs change, the home entity's responses will evolve accordingly.
  3. The home entity will perform the requested function and the results assigned a value by the resident.  Based on the values assigned, performance will be modified as the processes which produced the lower valued results are eliminated and those producing higher valued results are reintroduced to the population.
  4. The population of higher valued results that are kept by the home entity are then mated, reproduced, and utilized in the next iteration of the requested process.    Each requested process will be developed in this way.
  5. As the various populations of processes co-exist, they will co-evolve into an ecosystem and matrix network which will serve the resident.  These services and functions include but are not limited to: health concerns, communication, delivery services, temperature, lighting, cooking, recipies, security, timers/alarm clocks, research references, and tracking of fellow residents or pets. They can be automated and accessed from the home by the resident based on verbal or nonverbal commands and behavior.

Prior Art:

US4697242: Adaptive computing system capable of  learning and discovery
US6038556: Adaptive autonomous agent with verbal learning
US5299764: In-space servicing of spacecraft employing artificial life robotics
US5832469: Electronic system organized as a matrix network of functional cells

Patent Drawings