Patent Application
20060221769
The invention relates to a position estimation system for estimating a position of an object in a room, comprising:
an ultrasound emitter arranged to emit an ultrasound pulse;
an ultrasound receiver arranged to detect an ultrasound signal based on the pulse; and
a processor for processing the ultrasound signal to obtain the position.
The invention also relates to an apparatus for estimating a position of an object in a room comprising:
an ultrasound receiver arranged to detect an ultrasound signal based on an emitted ultrasound pulse; and
a processor for processing the ultrasound signal to obtain the position.
The invention also relates to a method of estimating a position of an object in a room, comprising:
emitting an ultrasound pulse;
detecting an ultrasound signal based on the pulse; and
processing the ultrasound signal to obtain the position.
The invention also relates to a computer program product enabling a processor to execute such a method.
In prior art such position estimation systems are typically realized as triangulation systems, such as the variant of U.S. Pat. No. 5,107,746. The known system variant comprises for estimating the position of an object in a plane at least three ultrasound emitters and an ultrasound receiver connected to the object. The system calculates the time it takes an ultrasound pulse emitted by each one of the emitters to travel to the receiver. If the three emitters are not collinear, the position of the object can be derived by means of the laws of geometry. More emitters may be present in a triangulation system for increased accuracy. Position is understood to be three dimensional position (x,y,z) e.g. in a coordinate system relative to a corner of the room, although in some systems only some of the coordinates may be measured.
It is a disadvantage of the known system variant that the possibilities of ultrasound position estimation are not exploited fully.
It is a first object of the invention to provide a system of the kind mentioned in the opening paragraph, which is, constructed to better exploit the possibilities of ultrasound object position estimation.
This first object is realized in that either the emitter or the receiver is attached to the object and the processor comprises evaluation means arranged to estimate the position of the object on the basis of the ultrasound signal in a first time interval, the first time interval comprising ultrasound energy of at least a first reflection of the pulse travelling towards a reflecting object on the reflecting object. Apart from the direct line of sight (LOS) transmission of the ultrasound pulse from the emitter to the receiver, in a room the pulse also reflects off objects, such as e.g. walls. The invention is based on an insight that a signature of these reflections can be used to estimate the location of the object. E.g. if an object is nearby a cupboard, the detected ultrasound signal may comprise a strong first peak corresponding to the direct transmission, a weak first reflection peak corresponding to a reflection off a wall towards the object and a relatively strong second reflection peak, corresponding to a reflection off the cupboard towards the object. If the same object is positioned in the middle of the room, the second reflection peak may firstly not be present or at least not so pronounced, and secondly it will arrive at a different time corresponding to the direct transmission peak. Hence the signature of peaks is characteristic of the object's position. The sound field in the room is as it were sampled at the position of the object. The first time interval is predetermined as to contain at least one peak corresponding to a reflection off an object. The temporal location of this peak may already be enough, in case e.g. a one-dimensional position along a narrow long table is to be determined via reflection of a ceiling of the room. In general more information can be retrieved from the location, height, etc. of any of the peaks, so the interval will typically be as large as to contain all meaningful reflections, while rejecting the reverberation tail, which will only confuse. Optimal values for this interval, determined e.g. from tests in rooms of different sizes and contents, may be pre-stored in the processor. The fact that reflections are used instead of or in addition to the direct transmission peak has a number of advantages. Firstly, instead of needing at least three, or for three-dimensional position estimation four emitters, a single emitter conveniently placed in the room is sufficient, since the different object positions have different corresponding ultrasound signal signatures. Because the system may be simple, more applications become feasible, for which classical systems would be too expensive. Secondly, triangulation needs clock synchronization, whereas some variants of the system according to the invention can work without a clock. Thirdly, the system is more robust in that it can tackle more easily the complex and possibly varying, in case the object of which the position is to be estimated is moving or being moved, configurations of objects in the room. There may be no direct line of sight for some object positions, since looking towards the object from the emitter when fixed in the room, the object may be behind another object. This is relevant for some applications. E.g. in an application to find a misplaced object, the misplaced object may typically be placed e.g. on a bookshelf in which case most of its sides are blocked by the sides of the shelf. However, even through a small crevice there will be ultrasound passing. An object may even be placed inside a cupboard, i.e. with the door closed, and the system may still work provided there is some ultrasound penetrating the cupboard. A related advantage is that the emitter, or receiver, may be hidden from direct sight, e.g. behind a panel or in the ceiling.
The emitter may be in a fixed position in the room and the receiver attached to the object. Vice versa, the system works in exactly the same way if the emitter is attached to the object and the receiver fixed in the room. This last option is very desirable for relatively cheap objects. Since a cheap object typically may not comprise the necessary processing capability to perform the position estimation, it may have to transmit the ultrasound signal detected to a processor in the room, e.g. by means of a bluetooth or other RF, IR etc. connection. However if the emitter is attached to the object, it may be comprised in a very simple configuration, e.g. a circuit incorporated in a label similar to what is known from theft protection or an attached ASIC, which emits an ultrasound pulse e.g. every 10 seconds. The receiver may then be comprised together with the processor in a module, which is put somewhere in the room by a user. Of course, both emitters and receivers may be attached to the object resp. fixed in the room, and as is known to a person skilled in the art of ultrasound transducers a transducer may function both as an emitter and a receiver.
There are also prior art object position estimation systems which position objects by ultrasound reflecting off the object into a receiver typically close to the emitter. These systems would not work very well in some of the applications for which the system according to the invention is intended, e.g. small objects can each be fitted with a label containing an emitter and hence reveal their position, but there would not be enough ultrasound reflected off them. And also in many systems with an isotropically emitting emitter or wide angle emitter, you would not know if the returning signal had reflected off the desired object or another object. Under pulse should be understood any of a variety of emitted signals, including e.g. a signal of short temporal duration. It is advantageous if the pulse signal has a sharp declining autocorrelation, as e.g. a Barker sequence or a noise signal. The working of the system according to the invention is described with a Dirac pulse or equivalently a very short duration emitted ultrasound signal.
An embodiment of the position estimation system comprises a storage means, arranged to store a first ultrasound signal template corresponding to a first position of the object and a second ultrasound signal template to a second position of the object, and the evaluation means are arranged to compute a first match by evaluating a predetermined match function with as arguments the ultrasound signal in the first time interval and the first template in a first template interval, and a second match by evaluating the predetermined match function with as arguments the ultrasound signal in the first time interval and the second template in a second template interval. For each position where the object may be put, e.g. on the coffee table, in the couch, on a cupboard, etc., there may be at least one template present in the storage means. Evaluation means are present, e.g. in the processor, to compare the ultrasound signal detected by the receiver with each of the different templates. Therefore a predetermined match function is evaluated, e.g. a correlation. Under match function should also be understood functions that are sometimes called error functions, like e.g. a mean squared error criterion, or in general any function that is capable of indicating a similarity between the detected ultrasound signal and the templates. The best matching template corresponds to the most likely position of the object. For increased robustness, several match functions can be evaluated, and their results can be combined in any way known in mathematics.
A further variant of the embodiment comprises collecting means, arranged to collect the first ultrasound signal template and the second ultrasound signal template on the basis of ultrasound signal measurements by the receiver, with the object positioned at a position in the room corresponding to the first resp. the second position. E.g., the receiver may be attached to the object, in which case it detects the local ultrasound field based on a pulse from one or more emitters typically fixed in the room, or an emitter attached to the object may also be used, in which case typically one or more receivers at fixed locations in the room detect an ultrasound field characteristic for a pulse emission from the particular position of the object in the room. The user may then position the object in a number of interesting positions, e.g. on the couch, and train the system by activating it. The object may e.g. comprise a microphone, which picks up the sound of a user's speech saying “couch”, and this speech may be stored by a second processor comprised in the object in a memory in the object. In an application to find a lost object, the speech saved for a particular estimated object position may then be reproduced with a loudspeaker attached to the object. No speech recognition is needed in this object embodiment, only a table in the memory linking uttered speech sentences and ultrasound signal characteristics. Alternatively, the user may e.g. enter position information, in a predetermined way, e.g. by means of user interface software, receiving as input room coordinates, or object names, etc., by means of a computer or other consumer apparatus, like a mobile phone, remote control, etc.
Another variant of the embodiment comprises simulation means, arranged to generate the first position and the second position of the object and calculate the first ultrasound signal template and the second ultrasound signal template on the basis of the pulse, predetermined characteristics of the room, and the first resp. the second hypothetical position. An ultrasound signal signature may also be calculated by simulation means instead of being measured. This saves the user the work of putting the object in different positions and training the system. Furthermore, positions may be estimated which were never trained. For this, the simulation means simulates how the pulse moves through the room and reflects off all objects. Predetermined characteristics of the room are needed. For simple configurations, these characteristics need not comprise much information. In an advantageous system configuration e.g., the emitter, or receiver, is positioned close to the ground and object positions are estimated by means of reflections off the ceiling. This largely avoids complicated reflections on the many objects that may be present in the room. Alternatively, a full simulation of all reflections may be performed by the simulation means. Therefore room characterizing means may be present to retrieve the room characteristics, e.g. a scanning ray may go through all angles in the room. Different simulation models may be used, e.g. ray tracing, or the image method. All simulated templates are subsequently stored in the storage means, e.g. chip memory or a removable disk, and during use of the system, a detected ultrasound signal is compared with all the stored templates as described above.
Another embodiment comprises optimization means, arranged to predict a predicted position of the object and evaluate a third ultrasound signal template on the basis of the pulse, predetermined characteristics of the room and the predicted position. Instead of prestoring templates, these templates may also be calculated by simulation on the fly. This may lead a faster system if the time to simulate a few templates, e.g. with a simple simulation system, is shorter than the time to compare all templates in the storage means. There may already be a good initial guess of the object position. The exact position may then be derived in a number of prediction steps, with any of the methods known form optimization theory, e.g. steepest descent. A manifold may be modeled in an N dimensional space having on its axes parameter values extracted from the signature, e.g. the positions and amplitudes of different peaks. The correct position may also be calculated directly, e.g. on the basis of the locations of the peaks of the reflections.
An advantageous variant of the embodiment is arranged to have the storage means contain more than one ultrasound signal template for a hypothetical position of the object, and the processor is arranged to estimate besides the position of the object a further parameter. E.g. if an object is lying on the table, the signature of its detected ultrasound signal may be different when a door is opened or closed. The further parameter being the extent to which the door is opened may or may not be actually used in an application of the system, but at least the estimation of the position is less error prone because the detected ultrasound signal is not yielding a better match with a template of an incorrect position. This may also be a variant of the system according to the invention itself, e.g. in the embodiment with the optimization means, and without storage means, for each predicted position a number of simulations may be performed with different parameters, e.g. room configurations.
In a further variant of the variant directly above, the processor is arranged to estimate as the further parameter an orientation of the object. If the object has a single receiver, ultrasound signals with different signatures may be detected depending on the orientation of the object. E.g., if the receiver is pointed towards the reflections of the cupboard, i.e. its angular sensitivity function is greatest in the direction of the cupboard reflection, a strong peak will be detected, but if the receiver is pointing towards another direction it will not be as pronounced. By identifying a template corresponding to a certain object orientation as best matching template, the object knows in what direction it is pointing.
If the object is a remote control e.g., it may send another command depending on whether it is pointing to e.g. a television or a stereo set. Orientation is understood as three angles in space, e.g. represented as pitch-roll-yaw (p, r, a). For an object like a remote control lying on a table only the estimation of the yaw a may be required.
In a multireceiver variant, the receiver or the emitter is attached to the object, and a second receiver respectively a second object emitter is attached to the object, and the processor is arranged to estimate the position of the object based on the ultrasound signal detectable by the receiver and a second ultrasound signal detectable by the second receiver, respectively based on the ultrasound signal detectable on the basis of the pulse emitted by the emitter and the second ultrasound signal detectable on the basis of a second pulse emitted by the second object emitter. E.g. if the object is a rectangular hexahedron, a receiver may be attached to each of the five surfaces not in contact with the ground, by combining the matches of ultrasound signals received from the different directions, a relatively robust position estimation may be achieved.
Similarly another multireceiver variant may easily estimate the orientation of the object. A single receiver, if non isotropic, may also estimate the orientation, but by analyzing the signals impinging on the object from different orientations this can be done more robustly or more accurately.
In a beam scanning variant, a plurality of emitters or receivers is comprised, arranged in at least one array, to create a number of pulses of different shapes and orientations respectively receive pulses impinging on the receivers from different orientations and in different solid angles. Techniques known in the state of the art of beamforming can be used to e.g. make an array of emitters on the object, which is capable of emitting a beam of narrow width in a predetermined direction. Different directions can be scanned through by changing phases of the signals to the different emitters. Similarly, an array of receivers on the object can be more sensitive to ultrasound signals approaching from particular directions. Beamforming is well known, but the application in the system according to the invention is based on the insight that more information can be obtained by using reflections arriving from an object under different angles. Furthermore, those directional reflections may be more characteristic, and only the most characteristic ones may be selected for matching, discarding the less characteristic ones, according to a predetermined characterizing power identification criterion, comprising e.g. the number of significant peaks, the interpeak amplitude variation etc.
In an object identifying variant, the emitter is attached to the object, and is arranged to emit the pulse comprising frequencies in at least one predetermined frequency band characteristic for the object. A first object may e.g. identify itself by emitting frequencies in a narrow band, e.g. several hertz upto several kHz, around 40 kHz, while a second object emits frequencies around 45 kHz. In this way, not only the position of the object is readily identified, but with minor modifications also the identity of the object is identified. Multifrequencyband codes may also be used, as well as other coding schemes, which use temporal, frequency, phase, etc. information.
In a tracking variant, tracking means are comprised, arranged to track along a trajectory the position of the object.
A system capable of tracking moving objects may have object guidance means to guide the object, e.g. a vacuum cleaner capable of automatically cleaning the room or a robot. The fact that the locations of all the other objects may be known, e.g. receivers on the four corners of a cupboard, is advantageous for guiding the vacuum cleaner around the objects, and contrary to the random cleaning strategy assuring that every square cm of the floor is cleaned. Furthermore, since a moving object has ultrasound signatures of different locations, a higher localization accuracy and/or robustness may be achieved. E.g. the processor may be arranged to combine a number of templates in the vicinity of a position, or the tracking means may be arranged to derive the current object position taking into account any combination of information on the trajectory, e.g. a likeliness of a trajectory, or signatures detected along the trajectory. E.g. by detecting a signature corresponding to a position in the vicinity of a table and at a later time instance a signature corresponding to a position in the vicinity of a cupboard, the processor is able to identify to which part of the room an object is moving, and in which part it can not reside.
An apparatus for estimating a position of an object in a room comprising:
an ultrasound receiver arranged to detect an ultrasound signal based on an emitted ultrasound pulse; and
a processor for processing the ultrasound signal to obtain the position, characterized in that the processor comprises evaluation means arranged to estimate the position of the object on the basis of the ultrasound signal in a first time interval, the first time interval comprising ultrasound energy of at least a first reflection of the pulse on a reflecting object is also disclosed.
It is a second object of the invention to provide a method of the kind mentioned in the opening paragraph, which is, constructed to better exploit the possibilities of ultrasound object position estimation.
The second object is realized in that:
the pulse is emitted from a position in a vicinity of the object or the ultrasound signal is detected in the vicinity of the object; and
the processing comprises an evaluation of the ultrasound signal in a first time interval comprising ultrasound energy of at least a first reflection on a reflecting object. This method i.e. its steps or actions may form part of different applications such as e.g. tracking an object along a production line or guiding a robot through a building.
A computer program product enabling a processor to execute the method is also disclosed.
These and other aspects of the position estimating system and apparatus according to the invention will be apparent from and elucidated with reference to the implementations and embodiments described hereinafter, and with reference to the accompanying drawings, which serve merely as non-limiting illustrations.
In the drawings:
a and 1b schematically show triangulation examples of prior art;
a schematically shows an ultrasound signal as detected by the receiver;
b schematically shows two ultrasound signal templates;
In these Figs. elements drawn dashed are optional, depending on the desired embodiment.
In
In
The position estimation system 220, 226, 222 is arranged to estimate a position of an object 224 in a room 200 by comprising an ultrasound emitter 220 arranged to emit an ultrasound pulse 201. If the emitter is isotropic, the same pulse shape travels to all directions in the room. The emitter may also be non isotropic, in which case a stronger pulse is emitted in some directions, and it may be so constructed that no pulse is emitted in other directions. By using the interference of pulses emitted by an array of emitters, a strongly directional pulse can be created, the direction of which can be controlled, i.e. different directions can be scanned. A similar system containing an array of microphones 602, 604, 606 and 608 is shown in
Advantageous positions for the emitter 220 may be the floor for off ceiling reflections or the ceiling for off floor, and off wall, reflections, in which case there are typically less high amplitude reflections off other objects. This may be less advantageous for supplying characteristic signatures but more advantageous for obtaining position estimates on the basis of the time of arrival of peaks corresponding to reflections, since these reflections will be more easily identified. Also the probability of LOS transmission is higher for an emitter/receiver on the ceiling. A fixed second emitter 250 and further emitters may be present for enhanced accuracy. Such a fixed emitter 250 should not be confused with a second object emitter 630 attached to the object 224. In the former case a receiver 228 on the object 224 is in the same position and orientation and samples as it were two different sound fields corresponding to exciting the room 200 with ultrasound from different directions. The ultrasound signal detected based on either sound field is enough for position estimation, however in case the signatures of some object 224 positions are so similar that confusion and hence erroneous position estimation may occur, a second sound field may resolve this issue. However the receiver 228 attached to the object 224 samples both sound fields at the same position and from the same direction. Attaching multiple emitters to the object 224, enables estimating the position of different parts of the object, and/or sending pulses through the room 200 in different directions, e.g. to ensure that there is at least one LOS transmission.
The pulse reflects off an object 232, such as e.g. a wall or a ceiling, as a first reflection 230 and travels towards an ultrasound receiver 226 arranged to detect an ultrasound signal 300. In
A processor 222 is also present, which comprises an evaluation means 223 arranged to estimate the position of the object 224 on the basis of a first time interval 320 of the ultrasound signal, the first time interval 318 comprising ultrasound energy of at least a first reflection 230 on a reflecting object. There are different algorithms possible to estimate the object position.
A first one is direct calculation. This is illustrated with
in which c is the speed of sound, xE and yE are the known position coordinates of the emitter, and xR and yR the position to be estimated of the object 224, in this case the duck 420.
This formula can be simplified to:
The only parameters to be known of the room in this algorithm is the height H of the room 200, and the height yR of the table. Furthermore also the position of the emitter (xE, yE) is known. In this variant of the system a synchronized clock for both the emitter and the receiver is required. However, in case the direct line of sight transmission 410 is present, it may be used to solve the problem of the unknown time of emitting the pulse 201. By using the ceiling reflection, this system may be conceived as a two emitter system as in
Another advantageous algorithm uses many reflections, at least those which are still characteristic enough, i.e. not the many small reflections in the reverberation tail. The principle exploited in this algorithm is that at least at each location of interest, e.g. for example on a shelf of the cupboard 208, on a coffee table 204 or on a couch 202, the ultrasound signal 300 detected by the receiver 226 during a time interval will have a characteristic signature of peaks. Hence the detected signature can be matched with a number of templates, having the typical signature of an ultrasound signal as it would be detected at a number of positions. Different algorithm combinations can be made with respect to the following variations.
A first variation is whether templates are prerecorded or calculated, i.e. simulated by means of a simulation means 227, e.g. software running on or a hardware circuit in the processor 222. In case they are prerecorded a user puts the object 224 of which the position is to be estimated, e.g. a remote control or a mobile phone, in different positions, e.g. five positions on the coffee table 204, some positions on different shelves of the cupboard 208, etc., and records the ultrasound signal, e.g. under different angles if the object is also rotated or multiple receivers resp. emitters are attached to the object. The user also marks a characteristic of the location, e.g. the name of the object the object 224 is on. All these measured templates are stored in a storage means. The evaluation means may match templates to obtain a position estimation in different ways. In a straightforward way, each template, corresponding to a particular object position, orientation, and possibly further parameters, such as whether a door 212 is open or closed, is matched with the ultrasound signal 300 detected during position estimation. This may be according to a variety of predetermined match functions, such as e.g. a coherence between a power density spectrum of the template and the ultrasound signal 300, a root mean square error, a cross correlation, a function of the peak positions, etc. Preferably a number of those criteria are combined in a single match function. A first ultrasound signal template 320 will match better with the detected ultrasound signal 300 than a second ultrasound signal template 330, since its peaks, first template peak 322 etc., are more similar to those of the detected ultrasound signal 300. In the schematic example this will be even more so if the match is restricted to a time interval 318, corresponding to a first template time interval 328 and a second template time interval 338.
A number of templates near to a particular position, e.g. all templates of positions on the coffee table 204, may be used in combination in the match criterion. E.g. a multidimensional space may be constructed on the basis of characteristics of the templates, e.g. time delay of the first, second, third, etc. peaks on a first, second, third, etc. axis, and then techniques from pattern recognition can be applied for characterizing the ultrasound signal 300. E.g. an N nearest neighbor classification may be used. Alternatively, the templates may also be calculated on the basis of the pulse and characteristics of the receiver and emitter and the room. E.g. a ray tracing algorithm may be used to simulate all the reflections and hence the signature at each position. The image method, which constructs pseudo sources for the reflections may also be used (see e.g. “J. B. Allen and D. A. Berkley: Image method for efficiently simulating small-room acoustics. Journal of the Acoustical Society of America, vol. 65, no 4, April 1979, pp. 943-950.”). The characteristics of the room may be acquired by the same system, e.g. in case emitter arrays are present, the room may be scanned, or via other means, e.g. the contents of room being a factory floor, or fume cupboard in which experiments are performed may be measured and entered in a different manner.
A second variation is whether or not the LOS transmission is present. If it is not present all positions in the room are possible candidates—or at least those within a radius corresponding to the first measured peak, since this could either be a LOS transmission or a reflection, in which case the object 224 is closer to the emitter 220 than the sphere radius, however if the LOS transmission is present only positions on a sphere with a radius corresponding to the time delay of the LOS peak are possible candidates. This decreases the probability of erroneous position estimation.
Instead of matching with predetermined or precalculated templates in the storage means 225, positions may be predicted on the fly and corresponding ultrasound templates may be calculated by an optimization means 229, e.g. software running on or dedicated hardware in the processor 222. E.g., an initial position may be known, especially if a moving object is tracked, and a new position may be predicted. By comparing the match with position deviations in different directions a better estimate may be obtained. If the match function is represented as an N dimensional manifold, different optimization strategies may be employed, such as steepest descent, but also other techniques may be employed such as e.g. a genetic algorithm. The peaks may also be modeled by parametric functions for regions of the room 200, e.g. Gaussians. The location of the mode and the height of the Gaussian is a function of the position and orientation of the object 224, and may be calculated for a number of positions and orientations given the reflections in the vicinity and characteristics of the receiver 226. E.g. if a first orientation a1 has a signature with a first peak of height h11 and a second peak of height h21, and a second orientation a2 has a signature with a first peak of height h12 and a second peak of height h22, an intermediate actual orientation a of the object may be estimated by calculating the differences of the actually measured peak heights and those of the stored or calculated signature corresponding to the first and second orientation. The signatures may be calculated roughly in a variant where speed is preferred over accuracy, e.g. in case only a few far apart positions need to be identified, and a hierarchical approach may be used where e.g. the region of the room is first determined by identifying characteristic reflection peaks, and a more accurate position is then determined on the basis of templates for that region.
The description of
The processor 222 may be comprised in a home PC, e.g. it may be realized by means of software supplied by a vendor of the system running on a general purpose processor, but alternatively it may be a dedicated device. It may even be comprised in the object 224.
A sophisticated variant of the object 224 such as a mobile phone may comprise collecting means 255 for collecting templates by detecting the ultrasound signal 300 at various object positions. The collecting means may comprise a user interface, such as a button, which can be pressed for activating the receiver, or a microphone for detecting the presence of and storing a spoken text. If the object 224 is simple, the collecting means 255 may be fixed in the room, e.g. attached to the processor 222. The simple object 224 may e.g. emit a pulse every few seconds or minutes at specific times, depending on among others a required temporal accuracy of position estimation and battery life, and the fixed collecting means 255 may be activated by voice command. A communication module 260 may also be attached to the object 224, e.g. a bluetooth link, which enables transmission of the detected ultrasound signal or time synchronization, compared to the speed of sound an electromagnetic signal travels instantaneously.
Different in house applications may benefit from the system, e.g. in a “find it” application, a remote control may be hidden behind a cushion on the couch. Upon estimating its position, the system may speak the word “couch” or show an icon of the remote control on an image of the room displayed on a display screen 210.
The object may determine its orientation a in the room, and depending on to which apparatus it is pointing (e.g. the object's IR remote control, or a camera, or a microphone is pointing), it may act different E.g. if it is a mobile phone, it may transmit to another mobile phone an image displayed on a television 206 or the display screen 210.
As illustrated in
If only a single isotropic emitter 220 is fixed in the room, care should be taken about its positioning. E.g. if it is placed in a corner of an empty room, confusions of positions mirrored with respect to the room diagonal may occur. A similar reasoning applies to a fixed receiver 226.
A second receiver 228, or emitter, may be attached to the object. Hence ultrasound signals detected from different directions may be used to more reliably estimate the position and/or orientation of the object 224. Each object or in general each emitter may emit a different signal, e.g. frequencies in one or more narrow spectral bands, so that each emitter is easily recognized. Alternatively they may emit signals at different reserved time intervals.
Tracking means 270 may be present, arranged to track along a trajectory the position of the object, e.g. as software running on the processor 222. E.g. Kalman filtering may be used to predict positions along the trajectory. The tracking means may supply candidate templates for matching or even select a different matching function depending on which part of the room 200 the object 224 is in.
a schematically shows a detected ultrasound signal 300. It is actually an electric signal representing an amplitude envelope of the ultrasound signal. This envelope arrives at the receiver 226 modulated on an ultrasound carrier. Typically filtering occurs to remove noise outside a frequency band of interest, demodulation, and analog digital conversion. There may also be matched filtering, taking into account characteristics of the emitter 220 and receiver 226. The peaks in this text are amplitude peaks and should not be confused with peaks in e.g. a correlation.
b schematically shows two example templates. It can be seen that a first ultrasound signal template 320 is very similar to the ultrasound signal 300 measured at a certain object 224 position, so the object 224 is probably at or very near to this positions, the differences between the two originating from e.g. measurement errors such as slightly different conditions in the room. The second template 330 is probably corresponding to an entirely different object 224 position. Note that t1 is the amount of time a LOS transmission takes to travel from the emitter to the receiver. A time reference from a clock is not always required in embodiments of the system according to the invention working with template matching, since regardless of the interval where there is little ultrasound energy detected before the LOS peak, the peaks will substantially match. Even when the first LOS peak is not present, the other peaks would still match relatively well.
Under computer program product should be understood any physical realization of a collection of commands enabling a processor—generic or special purpose, after a series of loading steps to get the commands into the processor, to execute any of the characteristic functions disclosed in the present invention. In particular the computer program product may be realized as data on a carrier such as e.g. a disk or tape, data present in a memory, data traveling over a network connection, wired or wireless, or program code on paper. Apart from program code, characteristic data required for the program may also be embodied as a computer program product.
It should be noted that the above-mentioned embodiments illustrate rather than limit the invention and that those skilled in the art are able to design alternatives, without departing from the scope of the claims. Apart from combinations of elements of the invention as combined in the claims, other combinations of the elements within the scope of the invention as perceived by one skilled in the art are covered by the invention. Any combination of elements can be realized in a single dedicated element
In particular, it should be clear to the skilled person that any combination of receivers and emitters fixed in the room and attached to the object can be employed. The system could even be used with two (moving or moveable) objects without fixed emitters or receivers to identify the relative position of the objects, provided that the room configuration and configuration of receivers and emitters is such that at least the positions of interest can be discriminated. E.g. a first transporting vehicle carrying supply goods in a factory may be following or avoiding another transporting vehicle, in which case a relative position of the two vehicles is to be determined. The storage means then contains templates of the ultrasound signals that can be detected in any relevant position, e.g. on a linear track the vehicle moves along, of the first vehicle, given the emitter attached to the second vehicle being in any relevant position. Knowledge on previous positions of the vehicles or other knowledge, e.g. a coarser position estimating system by means of beacons, photo cells or wires in the ground, can reduce the risk of getting erroneous position estimations. Any configuration can be used for position and/or angle estimation. In any of these systems, the characteristic ultrasound signal templates for each required position, orientation and further relevant parameters may be obtained by means of any measurement or calculation. In any of the previous systems, the estimations may be achieved by means of calculating properties of the detected ultrasound signal or by matching with any predetermined or on the fly determined combination of templates with any match or error function. And for any of the above described systems any feasible ultrasound signal may be emitted and additional processing may be done at the receiver side to obtain a better quality signal.
Any reference sign between parentheses in the claim is not intended for limiting the claim. The word “comprising” does not exclude the presence of elements or aspects not listed in a claim. The word “a” or “an” preceding an element does not exclude the presence of a plurality of such elements.
The invention can be implemented by means of hardware or by means of software running on a computer, and previously stored on a data carrier or transmitted over a signal transmission system.
| Number | Date | Country | Kind |
|---|---|---|---|
| 03101098.6 | Apr 2003 | EP | regional |
| Filing Document | Filing Date | Country | Kind | 371c Date |
|---|---|---|---|---|
| PCT/IB04/50465 | 4/16/2004 | WO | 10/18/2005 |
