Data processing: generic control systems or specific application – Specific application – apparatus or process – Robot control
Reexamination Certificate
2002-01-04
2003-05-06
Cuchlinski, Jr., William A. (Department: 3661)
Data processing: generic control systems or specific application
Specific application, apparatus or process
Robot control
C700S245000, C700S259000, C700S262000, C714S004110, C701S023000
Reexamination Certificate
active
06560512
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
This invention relates to selection of “cause” vector signals. The “cause” vector signals govern motors that control robotic limbs, optical sensors and audio generators and are used in navigating a robot through an “effect” vector space.
2. Related Art
A relational correlation sequencer includes a nodal map module and a sequence stepper module. Taken together, these modules generate (1) a nodal map and (2) a set of sequences for the control of one set of signals (denoted herein as p signals) based upon sequences of a second set of signals (denoted herein as q signals).
The nodal map shows a relationship between two sets of signals, wherein each set of signals is represented by a multi-dimensional vector field. The two multidimensional vector fields consist of a set of cause or control vectors (p vectors) and effect or sensor derived vectors (q vectors). Transitions in the nodal space are used to mirror transitions in a Euclidean or function space, and the control signals that cause such transitions. For example, a transition in the nodal map may represent a physical displacement in space and the control vector that causes such a displacement. The dimensionality of the nodal space may include one, two, three, four or more dimensions, preferably expressed as a mathematical construct in the form of a Kohonen self-organizing map.
The sequence stepper module generates a temporal relationship between the pq vector pairs. A sequence of specific pq vector pairs is generated by navigating through the nodal space, traversing the set of adjacent nodal points that define the sequence of pq vectors. Thus, given an initial nodal point defining a pq where t=0, and a final nodal point defining a q where t=n, the sequence stepper module will form a temporal sequence of pq that is defined by a navigational path between the initial nodal point and the final nodal point.
These relational correlation sequencers can be coupled together to form a hierarchy wherein each sequencer is responsible for a particular set of motors in a robot. For example, a first relational correlation sequencer may control movement of a robotic arm while a second relational correlation sequencer controls movement of robotic legs.
SUMMARY OF THE INVENTION
Numerous desired advantages can result from using multiple relational correlation sequences. For example, a relational correlation sequencer can store and receive data, thereby providing information that can be used to perform tasks on the basis of stored data, rather than data that is received in real time. Furthermore, a hierarchical set of relational correlation sequencers associated with a particular task can be used to “train” a robot to simultaneously perform other tasks that are related to the particular task. Lastly, a hierarchy of relational robotic correlation sequencers can provide memory and various sensory abilities that loosely emulate biological functions.
In a first aspect of the invention, a robotic device (and associated nodal map) is trained to perform a first task (for example, a task A) that involves a self location and identification task. For example, task A may involve moving a robotic finger so as to locate one or more pressure transducers used as tactile sensors that are uniformly distributed about the robotic body. On the basis of information relating to task A, the robotic device can learn to perform tasks B, C, D and others that relate to task A. In a preferred embodiment, this second set of tasks may involve inputs from different sensors (such as cameras mounted on a movable platform, sonar or radar sensors, auditory sensors, thermal sensors and so on) that are used to provide information to the device useful for the performance of the second set of tasks.
The combination of external sensory data and internal data enable the robot to operate in real time. The external sensory data is provided by the various sensors assigned to the robot. The internal data includes information such as the current flow and voltage associated with the motors and sensory systems, the charge in one or more batteries, the lubrication of moving parts and other information such as may reflect the internal state of the robot.
In a second aspect of the invention, a set of relational correlation sequencers may operate as a memory storage device that allows the robot to operate as a function of stored data, rather than real time internal and external data. For example, a correlation sequencer may be used to store all the data present in a street map. This data may be combined with other data inputs (for example, from one or more optical sensors) so as to provide intelligent navigation.
In a third aspect of the invention, a set of relational correlation sequencers add an auditory sound generating capability to the relational robotic controller.
REFERENCES:
patent: 4805107 (1989-02-01), Kieckhafer et al.
patent: 4816989 (1989-03-01), Finn et al.
patent: 4914657 (1990-04-01), Walter et al.
patent: 4933490 (1990-06-01), Iannella
patent: 4972415 (1990-11-01), Walter et al.
patent: 4980857 (1990-12-01), Walter et al.
Koca et al., Turbe space-time equalization of TCM with receiver diversity I: Symbol-by-symbol detection, 2000, IEEE, pp. 552-556.*
Frenkel et al., A dedicated DSP computation engine baased on VLSI vector signal processors, IEEE, 1989, pp. 2564-2568.*
Grenier, Wideband source location through frequency-dependent modeling, 1994, IEEE, pp. 1087-1096.*
Shinnaka, Proposition of new mathematical models with core loss factor for controlling AC motors, 1998, IEEE, pp. 297-302.
Rosen Alan
Rosen David
Cuchlinski Jr. William A.
Machine Consciousness, Inc.
Marc McDieunel
Swernofsky Law Group PC
LandOfFree
Relational robotic controller does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Relational robotic controller, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Relational robotic controller will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3061175