Abstract: A method includes receiving, at a mobile teleconferencing robot, a remote user input to alter a viewing state of a vision system of the robot. The vision system includes a forward imaging sensor arranged to capture a forward video feed, a right imaging sensor arranged to capture a right video feed, and a left imaging sensor arranged to capture a left video feed, each with respect to a forward drive direction of the mobile teleconferencing robot. The method includes altering the viewing state of the vision system by adjusting a tilt angle and/or a zoom level of the forward imaging sensor based on the remote user input and generating a combined video feed that provides an immersive peripheral view about the robot. The combined video feed is generated by combining the forward video feed with a portion of the right video feed and a portion of the left video feed.

Abstract: There is provided herein a modular video conferencing system. The system may include, for example, a main unit, a camera unit, and a docking station, each adapted to be removably attached to at least one of the others. The main unit may house a processor, memory, mass storage, and other hardware and software consistent with a video conferencing system. The main unit may include a camera adapter that electrically and mechanically connects the main unit to the camera unit. The camera unit may be directed to change pan, tilt, focus, zoom, and so forth through control signals transmitted from the main unit to the camera unit through the camera adapter. The camera unit may include a plurality of microphones that provide audio signals to the main unit through the camera adapter, and these audio signals may be used by the main unit to determine a location of an audio source relative to the camera unit so that the camera unit may be directed at the audio source.

Abstract: A public-key method of cryptographic key exchange using modular exponentiation in which memory, for storing pre-computed results, can be flexibly traded off against the computational complexity of key-exchange. In a typical embodiment, the invention performs key exchange by the method of Diffie-Hellman but with exponents having a constrained form such that by use of a small table of pre-computed powers of a user's public key, any possible shared secret key within the allowed set can be computed with many fewer modular multiplications than the number of bits of effective key-length thereby obtained. The table of pre-computed powers is transmitted as part of the key exchange protocol. The party in receipt of the table uses the pre-computed powers of the sender's public key to replace calculations that would otherwise need to be done at key-exchange time. The method allows a flexible trade-off between computation and table size.

Abstract: In a finite state machine having a plurality of registers, a stream of pseudo-random numbers is generated by a method having characteristic non-zero integral value N. Repeatedly, over a series of time steps, a set of N combined values is calculated by applying N non-linear combining functions to N input sets of values and registering each combined value in one of N registers. At any time step after N time steps each input set consists of combined values, except that one input set may include as one value the result of combining a combined value with a value from an input stream, and at least two input sets comprise only registered values from distinct registers.

Abstract: A finite state machine for producing a stream of pseudo-random numbers, having characteristic positive integral values S and T. The finite state machine has a w-bit non-linear combiner having an input and an output, a plurality of w-bit registers including at least a chained series of S registers and a chained series of T registers. The non-linear combiner receives as the input data from the last of the chained series of S registers and from the last of the chained series of T registers. The first of the chained series of T registers receives input from the last of the chained series of S registers. The first of the chained series of S registers receives input from the output of the non-linear combiner.

Abstract: A camera stand for directing a video camera at, alternately, a room scene and a document scene, where the document in the document scene is oriented to be read by a person seated in the room scene, featuring a supporting arm and a rotatable connection connecting the video camera to, and supporting the video camera at a supporting end of, the arm. The rotatable connection allows the video camera to be alternately rotated between a position directed substantially vertically down at the document scene and a position directed substantially horizontally at the room scene. The rotatable connection has a rotation axis inclined at substantially a forty-five degree angle to the vertical, and is connected to the arm so that the video camera rotates from viewing the room scene right-side-up to viewing the document scene right-side-up.

Abstract: Disclosed is an apparatus for synchronizing a first signal with a second signal comprising a plurality of delay means D.sub.i as i goes from 1 to N, where N is an integer, each delay means D.sub.i having an input I.sub.i and a delay output O.sub.i for delaying a signal received at the respective input I.sub.i by an increment .delta.t of time in supplying the delayed signal at the respective delay output O.sub.i. The first delay means D.sub.1 of the plurality of delay means is connected to receive the first signal at its input I.sub.1. Each of the other delay means D.sub.i, for i equal to 2 to N, are connected in series such that the respective input I.sub.i is connected to receive the delay output O.sub.i-1 of the preceding delay means D.sub.i-1. A plurality of latch means L.sub.i, as i goes from 1 to N, are connected to be clocked by the second signal. Each of the latch means L.sub.i latches the signal at the delay output O.sub. i respectively for each of the delay means D.sub.

Abstract: A conditional replenishment video data transmission system involves the transmission of updating digital data for clusters of consecutive picture elements, all or most of which show a change since the previous frame, in lines of a raster over an image to be transmitted. The lines of the raster are numbered cyclically in groups of 8, the numbers being part of line synchronizing code words, and the position along a line of the first element of each cluster is identified by a number. The data are transmitted at a constant rate from a transmitter buffer to a receiver buffer and an auxiliary data store records the addresses of the line synchronizing code words in the receiver buffer. The address of the receiver buffer read by a reading means can be corrected by reference to the address data in the auxiliary storage means if for example the line numbers do not follow in sequence or the element addresses are not in ascending order due to corruption during transmission.