GB2427319A - Intelligent mobile remote monitoring security system - Google Patents

Abstract

A mobile remote monitoring security system comprising video and other sensor input, an interface to a personal computer which contains software programs with embedded intelligence to enable signal processing and computer based decision making and a communication interface to transmit and receive signals using mobile phone and similar devices on the 3G GSM network. The signal inputs to the mobile remote monitoring security system comprise one or more web-cams supporting audio and video signal input, and a range of intruder detection sensors. Enhancement to the system would also include fire and gas sensing equipment. The system features intelligent software algorithms programmed into the personal computer which will use pattern recognition software to compare real-time images with those stored off-line and considered to represent the safe state. Based on this comparison and the processing of other sensor inputs an embedded computer program will make decisions as to whether or not a security threat exists. The user may be alerted when a security threat is detected. The police may be automatically alerted if no response is received from the user following a predetermined number of alerts.

Description

Report prepared by Dr. Goran / Bezanov PhD Jr and on heha//of MiG (.

onsu//ing Lid, (Company regis/rat ion Number O54O6O/ litle of the Invention Intelligent Mobile Remote Monitoring Security System I Backgrou of the Invention This invention relates to a mobile remote monitoring Security systems which embodies artificial intelligence (AT) capabilities for decision support in the event of a security alert within the area being monitored The invention comprises an intelligent central monitoriiig station with web-cam imaging capabilities which is remotely linked to a mobile phone and other similar devices. The central station further incorporates a personal computer running software algorithms for image Processing, decision support and suitable hardware interfce to enable communications via the GSM network. Please refer to figure 1 for the overall layout of the system. Where in the event of an intrusion being triggered via Processing (using the AT algorithms) anywhere in the world the device could receive live video feed, using the existing 3G technology GSM mobile phone network sent via the multimedia messaging service (MMS), to alert the user of an intrusion. The user has the option to reset the system if the received alert is deemed false, or to instruct the systeni to proceed with the process of contacting the local police authority with whom a previous agreement to deal with security alerts, has been arranged.

Current state of the telecommunications industry offers the ability to communicate data and voice using wireless techniques across the globe. Numerous techniques exist to digitise analogue signals, package these and if necessary compress and encrypt the signals for transmission A range of protocols are incoorated to ensure communication takes place in a controlled manner and additional methods are available for ensuring the correctness of transmitted data.

There follows a brief description of the technologies that are relevant to this invention, namely, GSM, MMS and Al techniques in pattern recognition 1.1 Global System for Mobile Communications GSM [I] This invention aims to utilise the existing GSM telecommunication infrastructure to provide the framework for transmitting and receiving data between an intruder detection system and the remote user. In this context the term GSM means the GSM standard and protocols in the frequency spectrum around 900MHz. There is also DCSI800 - GSM protocols hut at different air frequencies around 1800 MHz - and in the United States, where spectrum for Personal Communication Services (PCS) is at 1 900MHz. However, although the physical frequencies used differ, the protocols and architecture remain the same and this invention applies equally to all GSM standards.

GSM is continually evolving and at present the GSM family of technologies consists of existing GSM, General Packet Radio Service (GPRS), Enhanced Data rates for GSM Evolution (EDGE) and third generation GSM services (3GSM) based on W-CDMA and HSDPA access technologies. [1} Together these technologies underpin the GSM platform to which this invention relates.

Author 0/this report Dr. Goran I Bezanov PhD This report is prepared under instruction fr)m Mr. John Hendric.on by the Author in accordance with the agreeme signed between them Ofl 18/05/006 Report prepared by Dr. Goran I Be:anov PhD Jr and on behalf of M1G Consulting Lid, (Company registratjo Number 054060/8) It is also woiTh noting that the current state of 3G technologies is soon to be replaced with infrastructure It is anticipated that 4G would refer to whatever is deployed in the 2010-20 15 period, assuming 30 deployme11 spans the 2000-2009 period. Typically, this means a new air-interface with higher data rates and some changes in the way data transport is handled end.

Ideally, 40 would provide users with Ofl demand high quality video and audio. It is anticipated that video transmission will be the main differences between 4G and 3G. 40 may use OFDM (Orthogonal Frequency Division Multiplexing), and also OFDMA (Ohogonai Frequency Division Multiple Access) to better allocate network resources to multiple users. 40 devices may use SDR (Softwaredefifled radio) receivers which allows for better use of available bandwidth as well as making use of multiple channels simultaneously.

Additionally, unlike the 3G networks which are a combination of circuit switched and packet switched networks, 40 will be based on packet switching only. This will allow low-latency data transmission [2] Thus it is noted that the evolution into 40 will fuher enhance the practical usefulness and application of this invention.

1.2 GSM System architecture [3] The inctional architecture of a GSM system can be broadly divided into the mobile station, the base station subsystem, and the network subsystem Each subsystem is comprised of functional entities which communicate through the various interfaces using specified protocols This invention intends to use the GSM system architecture as it exists without any modifications being required. The components of the architecture are briefly described below.

1.2.1 Mobile Station The mobile station in GSM is really two distinct entities. The actual hardwaje is the mobile equipment, which is anonymous. The subscriber information, which includes a unique identifier called the International Mobile Subscriber Identity (IMSI), is stored in the Subscriber Identity Module (SIM), implemented as a smart card. By inseing the SIM card in any GSM mobile equipment, the user is able to make and receive calls at that terminal and receive other subscribed services. By decoupling subscriber information from a specific terminal, personal mobility is provided to GSM users. In this invention it is assumed that the communication devices in the central station and the remote locations are devices of this type (i.e. mobile phone, GSM modem etc) 1.2.2 Base Station Subsystem The Base Station Subsystem iS composed of two pans, the Base Transceiver Station (BTS) and the Base Station Controller (BCS). The BTS houses the radio transceivers that define a cell and handles the radio (Urn) intertce protocols with the mobile station. Due Author (?!/hjs report Dr. (loran I Bezano' PhD - This report is prepared under instruct ion from Mr. John Hendrickson by the Author in accordance with the agreement signed between them on 18/05/006 Report prepared by Dr. Goran I Bezanov PhD Jbr and on behalf of 1G Consu/tJng Lid, ((.ompany registraj Number 05406018) to the Potentially large number of BTSs, the requirem5 for a BTS are ggedness reliability, portability, and minimum cost.

The Base Station Controller (BSC) manages the radio resources for one or more BTSs, across the Ahis interface. It manages the radio interface channels (setup, teardown frequency hopping, etc.) as well as hand?vers In this invention it is assumed that the mobile services provider (i.e. Vodafbe, Orange e(c) will supply the base station and related services 1.2.3 Network Subsystem [3] The central component of the Network Subsystem is the Mobile services Switching Centre (MSC). It acts like a nomial switching node of the PSTN or ISDN, and in addition provides all the functionality needed to handle a mobile subscriber, including registratjo authentication location updating, inter-MSC handovers, and call routing to a roaming subscriber. These services are provided in conjunction with four intelligent databases, which together with the MSC form the Network Subsystem The MSC also provides the connection to the public fixed networks.

The home Location Register (HLR) contains all the administrative information of each subscriber registered in the corresponding GSM network, along with the current location of the subscriber. The location assists in routing incoming calls to the mobile, and is typically the SS7 address of the visited MSC. There is logically one HLR per GSM network, although it may be implemented as a distributed database.

The Visitor Location Register contains selected administrative information from the HLR, necessary for call control and provision of the subscribed services, for each mobile currently located in the geographical area controlled by the VLR. Although the VLR can be implemented as an independent unit, to date all manufacturers of switching equipment implement the VLR together with the MSC, so that the geographical area controlled by the M5C corresponds to that controlled by the VLR. The proximity of the VLR information to the MSC speeds up access to information that the MSC requires during a call.

The other two registers are used for authentication and security purposes. The Equipment Identity Register (EIR) is a database that contains a list of all valid mobile equipment on the network, where each mobile equipment i5 identified by its International Mobile Equipment Identity (IMEI). An IMEI is marked as invalid if it has been reported stolen or is not type approved. The Authentication Centre (AuC) is a protected database that stores a copy of the secret key stored in each subscriber's SIM card, used for authentication and ciphering on the radio channel.

For the purposes of this invention it is assumed that the network and related services are provided by the mobile services provider (i.e. Vodafone, Orange etc) and that all devices are suitably licensed to use the services.

Author (?tth/s' report Dr. Goran / Bezanov PhD This report is prepared under instruct ion from Mr. John Hendrjck,s.on by the Author in accordance with the agreemen, signed between them on 1 8/05/006 Report prepared by Dr. Goran 1 Bezanov PhD Jbr and on behalf of MiG Consulting Lid, (Company registratjo Number 05406018) 1.3 Multimedia Messaging Service (MMS) [4] For the purpose of alerting the user to intervene with the mobile security system it is intended that the existing 30 MMS service will be utilised to send video images. In short, MMS is a standard in mobile messaging. Like SMS (Short Messaging Service), MMS is a way to send a message from one mobile to another. The difference is that MMS can include not just text, but also Sound, images and video. It is also possible to send MMS messages from a mobile phone to an email address.

Formats that can he embedded with in MMS include: * Text (firmatted with fonts, colours, etc) * Images (JPEG, GIF format) * Audio (MP3, MIDI) * Video (MPEG) Images could be downloaded from WAP sites, selected from a menu within the phone, or could even be photoken using a built-in camera if the phone has one (e.g. the Nokia') 7650, Sony EricssoT0 etc).

One of the main practical differences between MMS and SMS is that whilst SMS messages are limited to 160 bytes, an MMS message has no size limit and could be many Kbytes in size, or even larger. MMS requires a third generation (3G) network to enable such large messages to be delivered although smaller messages can be sent even with second generation networks using GPRS. {4j In this invention it is intended to store the video images on the PC in MPEG format and to send these in MMS format when it is appropriate to do so. It is noted that when bandwidtl restrictions apply it is equally possible to store and transmit images in JPEG format.

1.4 Al techniques in Pattern recognitlo [5] Pattern recognition systems are widely used in the present day Al applications, and this invention merely aims to utilise some of the available tools in the context of intelligent remote monitoring systems.

Briefly stated, pattern recognition is a field within the area of machine learning.

Alternatively it can be defined as "the act of taking in raw data and taking an action based on the category of the data" [6]. As such, it is a collection of methods for supervised learning.

Pattern recognition aims to classify data (patterns) based on either a priori knowledge or on statistical information extracted from the patterns In the context of this invention the digitised images from web- cams are data patterns. Thus the pattern recognition concept can be readily applied to recognise the patterns that are of interest.

Author qithis report Dr. Goran I Bezanov PhD This report is prepared under instruction from Mr. John Hendrickson by the Author in accordance with the agreement signed between them on /8/05/006 Report prepared by Dr. Goran I Bezanov PhD for and on heha//oTMIG Consulting Lid, (Company registration Number 054060/8) A complete pattern recognition system Consists of sensors (i.e. web-cams) that gather the observations (data) to be classified or described; a feature extraction mechanism that computes numeric or symbolic information from the observations (i.e software written in a computer programming language (i.e. Visualflasjc 6. 0, Java, C++ etc); and a classification or description scheme that does the actual job of classifying or describing observations, relying on the extracted features (i.e. Artificial Neural network [13]) The classification or description scheme is usually based on the availability of a set of patterns that have already been classified or described. l'his set of patterns is termed the training set and the resulting learning strategy is characterised as supervised learning.

The classification or description scheme usually uses one of the following approaches: statistical (or decision theoretic), syntactic (or structural). Statistical pattern recognition is based on statistical characterisations of patterns, assuming that the patterns are generated by a probabilistic system. Structural pattern recognition is based on the structural interrelationships of features. A wide range of algorithms can be applied for pattern recognition, from very simple Bayesian classifiers to much more powerful neural networks. This invention proposes to use Artificial Neural Networks (ANN) and numerous vendors provide the tools to implement ANN. [l3],[14J. The invention intends to incorporate such a tool into the decisioti making process as depicted in flow-charts of figure 2 and 3.

At present, pattern recognition applications include automatic speech recognition, classification of text into several categories (e.g. spam/non-spam email messages), the automatic recognition of handwritten postal codes on postal envelopes, or the automatic recognition of images of human faces. The last two examples form the subtopic image analysis of pattern recognition that deals with digital images as input to pattern recognition systems. [5J It is also worth mentioning that Neuro Dimensions] offer NN solutions in a range of applications to include image processing. They use a Principal Component Analysis (PCA) type network to identify characters in images or video feeds. [14] Author o/thiv report Dr. Goran I Be:anoy PhD This report i.s prepared under in.structin/;.om Mr. .John Hendrickson by the Author in accordance wi/h the agreement signed between them on / 8/05/006 Report prepared by Dr. Goran I Bezanov PhD Jr and on heha//of MIG Consulting Ltd, (Company registration Number 054060/8)

2 Description of Related Art

Home security monitoring and surveillance systems are abundant today and in many cases these are hard-wired and do not support remote monitoring features.

Numerous inventors have addressed the need for remote monitoring by filing patents in the general area of remote surveillance. Following is a brief survey of a few patents relevant to this invention Generally speaking, remote communications of all types is a means to communicate information over long distances. In the context of a remote monitoring system the signal carriers and the protocols used in communicating the information are already available, most protocols are open (i.e. not Proprietary) and can be freely used by anyone.

In all cases of long distance communications a signal needs to be suitably prepared for transmission on the carrier wave (i.e. modulated) at the transmitting end and the transmitted signal needs to be demodulated at the receiving end. The medium for communication can be wired tbr example, Takashi [7J proposes utilising the existing electricity networks to transmit the modulated surveillance signal, or transmission could he wireless utilising radio frequency (RF) communication For example, Turner proposes to use RF caiers and sub- carriers to modulate the surveillance data for transmission [8], while Gonzales [9] goes a step further to file a patent for a very broad range of communications media that can he utilised in the applicatio of a remote monitoring system, including ATM, OSM, (iPRS, IISCSD and other 3G wireless networks. Tanaka [10] uses the mobile phone network to send the image to the remote user who will then decide whether to send this image to the police for attention.

It is noted here that none of the researched patents propose the use of AL techniques in data Processing and decision making for security monitoring systems.

Surveillance systems may incorporate a range of devices to include motion sensors, cameras, pressure Sensing pads, switches etc. Additionally Takashi [7] proposes that fire and gas leak detection sensors can he incorporated in the general context of surveillance and security monitoring of property. In all sensing instrumentation there is a need to process the signal in some way in order to render it useful to the user. The Processing of this signal can range from simply generating a trigger to indicate activation of the sensor (i.e smoke detector activated) to computationally intensive Processing of Video Images for transmission International standards for digital signal Processing (DSP) of video images are in place to ensure that these signals are rendered suitable for transmission.

Gonzales [9J identifies some of the suitable aIgorithis for DSP applications. All of these algorithms are currently available for anyone to implement in providing DSP solutions.

Gonzales also states that an ASIC chip can be designed to implement DSP Solutions, which of course is a standard means of implementing mass produced devices.

Having transmitted and processed the necessary data in a remote monitoring system application it is necessary to make some decisions as regards what action to take. Sending video images to the remote user will enable them to contact the relevant authority i.e. the local police as suggested by Tanaka 110].

Author oJthis report Dr. (loran I Be:anov PhD This report is prepared under instruction from Mr., John Hendrickson by the Author in accordance with the agreement sgned he/wee,i them on 18/05/006 Report prepared by Dr. Goran I Bezanov PhD for and on behalf of M1G Consulting Lid, (Company registration Number 054060/8) However, some problems are immediately identifiable here on the following issues: I. How often does a false alarm occur? This could present a problem if the remote user is inundated with monitoring system calls.

2. What happens if the remote user is not available to make the decision.

Author (?tthis report Dr. Goran I Bezanov PhD This report is prepared under instruction from Mr. .John Hendrickson by the Author in accordance with the agreement signed between them on /8/05/006 Report prepared by Dr. Goran / Be:anov PhD Jir and on behalf of MIG Consulting Lid, (Conpanv registration Number 05406018)

3 Summary of the invention

The problems that can arise from the current implementations (suggestions) in regard to remote monitoring systems as identified above can be overcome by incorporating artificial intelligence (Al) techniques in the processing of signals so that the remote user is alerted only after the Al algorithms have determined that user intervention is required.

In this invention it is proposed to use the existing mobile telecommunications infrastructure, including technology, devices, comrnunjcatiois media, protocols and DSP algorithms, to communicate the inlorniation between the monitoring site and the remote user. User intervention is simply to send one of two mutually exclusive signals back to the monitoring site. The acknowledge (ACK) signal will instruct the monitoring site to take a predefined action in response to a confirmed alarm, and a reset (RST) signal will cancel the alert and reset the monitoring station to the ready state. In this manner user intervention is simple and it is maintained at the highest level of decision making.

It is noted here that for alternative applications, such as monitoring the safety and situation of occupants who may need attentions (i.e. sickly relatives etc.), the signals sent from the user can he more elaborate. (i.e. switching between web-cam - rooms, alerting a nurse, requesting a response signal from occupants etc..).

The invention embodies Al techniques for processing the video sensor data in order to establish if a genuine security threat exists. Thus, the decision making process is rendered hierarchical in that the Al software perfonns all the local decision making, based on the available video sensor data, hut the decision whether to act on an alert signal rests with the remote user. In this manner problem I identified above has been solved. Which is to say that an MMS message will be sent to the remote user only after the Al pattern recognition algorithms have been applied. However, problem 2 identified above still remains. Namely, what happens if the remote user is unable to respond to an alert. To deal with this the Al algorithm incorporates a counter which is incremented every time that an alert is sent to the remote user. If the counter reaches a level that is set as a maximum (i.e. 3 alerts), the Al algorithm acts upon this as though the ACK signal were sent from the remote user. That is to say, a predefined action is taken in response to a confirmed alarm.

The degree to which an Al software solution goes in order to determine the verity of the security threat will depend on the number and the arrangement of sensing equipment. In this invention it is proposed to use pattern recognition tools to pre-process the digital images received from the camera sensing equipment. In particular it is proposed to use Artificial Neural Networks (ANN) for this purpose.

In the flrst instance this invention proposes to process only the video images for the purpose of security monitoring In other applications of this invention, the input from additional sensing equipment can be processed to complement image Processing an enhance the process of decision making to determine the security state of the system being monitored. (see figure 4).

Author (?tthis report Dr. Goran I &Zanov PhD This report is prepared under instruct ion from Mr. J()hn hendrickson by the Author in accordance with the agreement signed between them on 18/05/006 Report prepared by Dr. Goran I Bezanov PhD for and on behalf of MiG (onsufIiF7g Lid, (C'onpony registratjo Number 05406018)

4 Brief Description of the drawings

Figure 1 shows the overall layout of the intelligent mobile remote security system The systems comprises three sections, namely Video Monitoring system (1), Image Processing and decision support system (2) and GSM communication system (3) There are also interface components (hardware and software), between the Video Monitoring system (I) and the image Processing and decision support system (2) as well as an interface between the mage Processing and decision support system (2) and the GSM communication system (3). These interfaces enable the integration of the respective systems into the complete system for intelligent mobile and remote security monitoring.

4.1 Video monitoring System (1) This system comprises of a number (N) of web-cams. These are standard USB web-cajs with 640x480 pixel resolution (VGA), used for real-time video recording and Interuet applications such as chatting and video conferencing The cameras are equipped with snapshot button and integrated pre-ampljfied microphone (example item Trust 320 Spacecani [II] This web-camp1ies digitised video images through the EJSB port to any PC running Windows Similar devices are available for other operating systems (i.e et). The cameras are inexpensive and at the time of' writing they can be purchased for approximately 10 each. Interface of the cameras to the PC is a standard USR 2.0 interface The video monitoring subsystem feeds live images to the image Processing system decision support system (2) Figure 4 shows an enhancement to the video monitoring where additional sensors can be used to monitor intrusion. These sensors can include audio, pressure sensing, door switches and similar intrusion detection sensing equipment. Additional sensing such as fire and gas leak detection can also be incorporated 4.2 Image Processing and decision suppo,1 system (2) This system comprises a PC running Windows NT (2000, XP etc) operating systems. l'he PC is equipped with a suitable number of USB ports to accommodate the number (N) of web-cam feeds. The PC iS configured to support the web-cams by installing the suitable drivers according to manufacturers recommendations Processi of the images is perfore in the software. Microso messenger (MSN) and Yahoessenger as well as Skyped others all provide the software that enables live- video images to he transmitted in real-time. The image Processing and decision support system (2) intends Using similar software tools to process the images from the web-cam.

This tools form a part of the interface between the Video Monitoring system (I) and the image processing and decision support system (2) as shown in figure 1.

These images are then stored in a run-time buffer in the PC (or dedicated DSP hardware) in order to limit usage of disk space. The size of this buffer depends on the level of security that the system is required to provide. Thus the images that are supplied by the Author (?!thi report Dr. Goran I Bezanm, PhD This report is prepared under instruction tram Mr. John IIendrickon by the Author in accordance with the agreement signed between them on /8/05/006 Report prepared by Dr. Goran 1 Be:anov PhD Jir and on behalf of MIG Consulting Ltd, (Compan registra/jo Number 05406018) web-cam through the application programs on the PC are stored in the cache memory of the buffer.

A dedicated computer program comprising a number of functions is written in Visual Basic or C++ perfbrms the processing of these images in order to implement the Intelligent decision making process.

Figure 2 shows the basic flow chart of the main program implementing the system for image Processing and decision suppo (2). The program initialises the counter C to zero.

This counter is used to action a response in the event that the remote user is not able to acknowledge the alarm sent to them. For example, if after three attempts (ie CMAX3) to alert the remote user, no response from them is obtained, the system sends an a1e to the police without remote user intervention It is presumed here that the police have been prepared and that suitable agreemen5 have been made to allow for this process to happen.

The next action is to initialise the flag called LEARN which is later used to detennine if the dthult images have been stored. The LEARN 0 is the first setting of the flag which is used to determine if the images that are considered to he the SAFE state' of the system have been stored. The priming of the system will store the SAFE state' images that will be used for comparison with real-time images. After these images have been stored (i.e. learned). The LEARN flag will be set to I and no lrther learning will be performed until the complete system is reset. The program flow chad shown in figure 3 shows that the LEARN=J is performed after the A algorithm has completed the LEARN Image Process, and after the learned images have been stored.

Returning to the program flow chap of figure 2, the flow proceeds with receiving real- time digital video input from web-cams I -N. Where N is the total number of web-cams used on the central monitoring site. These are the images stored in the aforementioned runiling buffer. After the images have been acquired, the loop proceeds with implementing a pattern recognition algorithm shown in figure 3. In this algorithm an Artificial Neural Network is used to perform a quick compasisoii between the SAFE image and the current image for any camera 1-N. The detail of figure 3program flow will be discussed later, hut for now let us assume that if a change in the pattern is detected for any camera image (1-N), the program of figure 3 will send a TRIGGER signal. Returning to the program of figure 2, the TRIGGER signal is sampled. If it is not detected the program loops to web- cam input and the whole process is repeated until TRIGGER has been detected. When the program of figure 2 detects a trigger, it proceeds with incrementing the counter C. Thus a running count is maintained to record how many TRIG(ER events have occurred. The program proceeds to check if the maximum allowed number of TRIGGER signals have been reached. The maximum number can be set by the user. Three is a suitable choice for arguments sake. At this point in the flow- chad there are two options to consider. Firstly, if the number of TRIGGER events are three the system will contact the police directly. As soon as the police have been contacted the counter C will he reset to 0 and the program will return to inputting VII)E() IN signals from web-cams I -N. It is assumed that at some point the police will Author ?t this report Dr. (ioran I Bezanov PhD This report is prepared under instruction from Mr. . John Hendrickson by the Author in accordance with the agreemen, signed between them Ofl 18/05/006 to Report prepared by Dr. Goran / Bezanov PhD for and on hehalto/MiG C.'onsu/ting Lid, (Company regis/rag/ Number 05406018) visit the premises and the system will be returned to SAFE state. Until such a visit has happened the program will continue with the monitoring and decision making process, incrementitig the counter C every time the TRIGGER signal has been detected. The second option is when the number of accumulated TRIGGER signals is less than three. In this event an MMS is sent to the remote user who can view the video images sent. The user then decides whether action is required. The choice is whether to ignore the TRIGGER because the remote user considers that no further action is needed in which case the user will send the reset (RST) signal back to the monitoring system.

Alternatively the user may decide to acknowledge that the alann is valid and send the acknowledge (ACK) signal back to the intelligent monitoring system. If such a signal is sent the Counter C is set at the maximum value (in this case three) and the program returns to process the video images. When the program next reaches the decision to call the police, counter will exceed the maximum value and the police will be contacted.

Thus it is shown in figure 2 program flow chart that the value of counter C determines whether the user or the police should be alerted. The program Continues alerting in this manner until the site is visited by the police and the system reset to SAFE state.

Looking at the program flow chart of figure 3 more details about the pattern recognition section of the system can he obtained. The starting point of the program in figure 3 is to determine if the SAFE images have been stored, This is done by examining the state if the LEARN flag. If it is zero, as at initialisation, then the program will proceed with learning the images using an Artificial Neural Network (A). (i.e NeuroSolutions 5.0 114]). This will be done for images of all web-cams I -N. In an unoccupied space a still image can he used, hut in an occupied space a sequence of video frames can be processed instead.

The principle of using the ANN to compare images will apply equally to a still frame and to a sequence of video frames. The only difference is the Processing time and memory requirern5 of the system. Having learned the images, the program proceeds with storing these so that they can be used for comparison with real-lime images. At the next step LEARN flag is set to I to indicate that all the images have been learned and that for the time being, no further learning is required. The program proceeds with real-time comparison of images received from web-cams I -N (CAMS N) with the stored images for respective web-cams I -N (STORE N). The choice of web-cam N is set to I to select the first web-cam CAM I, and for this webcam a comparison is made with the stored image for the same camera STORE 1. The comparison is made by the neural network algorithm (A) allowing for a quick decision on the pattern match. If the match is made, which is to say that the images match, the next web cam is set for comparison, that is N is incremented. Before returning to image comparison for web- cam N, the program checks if the current comparison was using the last web-cam in the system, i.e. NMAX If this is the case, program returns to setting Nl, which is to say the comparison will start from the first web- cain in the system. It on the other hand the current comparison is not for the last web-cam, the program will return to examine the next web-cam in the sequence. The program of figure 3 therefore performs the comparison of images and signals the TRIGGER when the comparison yields a difference in the image pattern Author (?f this report Dr. Goran I Bezanov PhD This report is prepared under instruction from r..John Hendrickson by the Author in accordance wi/h the agreemen/ signed between them on /8/05/006 Report prepared by Dr. Goran / Bezanov PhD Jr and on behalffMiG Consulting Ltd. ((ompany registration Number 05406018) between the stored image and the real-time image for each web-cam. This TRIGGER signal is returned to the program of figure 2 where it is processed further.

Figure 4 shows the layout of an enhanced arrangement for the Image processing and decision support system (2). In this figure additional software is shown that will process signals other than video images. These signals could be from pressure pads, door switches and other sensors used in modern intruder detection systems. The software used to process these signals would be based on a set of rules which when applied will determine if the system state is SAFE or otherwise. The algorithms used in this section can be used to complement the pattern recognition software so that a more complete information of the state of the monitoring system is available. For example, if the patter recognition system recognises a TRIGGER and at the same time, pressure sensors and door switches indicate an intruder presence, then it can be inferred that the system security has been violated and that police intervention is inimediately required. The flow chart for such a modification has not been drawn but it is noted that it could easily be incorporated into the proposed system 4.3 GSM communication system (3) The GSM communication system (3) shown in figure 1 is based on the 3G GSM devices and technology. This system is instructed by the image Processing and decision Support system (2) to send an MMS to the user or to send a message to the police. A suitable interface between these two systems is required to enable GSM communications (i.e. IN4MA GSM modem [12]) In order for this to he done an interface between the image Processing and decision support system (2) and the GSM communication system (3) must be established. The GSM systems architecture as describe in the background section of this report relies on three components namely, the mobile station, base station and the network system. The proposed intelligent monitoring system intends to use the existing GSM infrastructure as described in the background information and other referenced text.

Thus, this report proposes using a communication system comprising of the devices described in GSM architecture. Thus it will comprise the receiver sp which at the remote user end is a mobile phone with MMS capability (i. e. Motoro V600 etc). At the system monitoring end the PC needs to be enabled for communication using GSM. This requires that a hardware is installed in or interfaced to the PC to support GSM.

Numerous devices exist for this puose. For example IN4MA Poweelectrics offer a GSM alarm dialler [12] that will send text messages directly to your mobile phone. Their Pocket GSM alarm dialler has 4 digital inputs, so when one changes state a text message is sent directly to your mobile phone. There are two digital outputs also so it is possible to send text messages to the Pocket GSM to switch things on and off This device utilizes an 8 bit 40Mhz microcontroller for alarm processing coupled to a GSM wireless module.

The unit can be programmed directly from a mobile phone, so to make changes it is not necessary to travel to site. The unit uses a standard cellular phone SEM card readily available from any mobile phone service operator, which is easily inserted by the user.

Author of this report Dr. Goran I Bezanov PhD This report is prepared under instruction /iom Mr/oh,7 Hendrickson h' the Author in accordance with the agreement signed between them on 18/05/006 Report prepared by Dr. Goran I Bezanov PhD J?r and on hehaUofMjG (.onsuiting Lid, (Company registration Number 05406018) The same manufacturer also provide a larger unit with more Iatures. Namely the, IN4MA GSM remote telemetry unit (GSM RTIJ) which communicates across the mobile phone networks to remotely monitor machinery, process, or stocks etc. It is noted here that these are only examples of a multitude of currently available GSM devices in order to show that the necessary technology is in place to suppo the idea behind this invention.

For the pu1)ses of this report, it is not required to delve into the detailed architecture of such a device since it is not intended to modify this technology in any way, simply to use it.

Thus, having established GSM communications between the monitoring system, the remote user and police messages can be sent between these points. It is possible to alert the police via an e-mail message rather than sending an MMS Video message, but these decisions do not impact on the overall description of system capabilities.

Author of this report Dr. Goran I Be:anov PhD This report is prepared under instruction from Mr. John /-/endrickson by the Author in accordance with the agreement signed between them on 1 8/05/006 Report prepared by Dr. Goran I Bezanov PhD for and on behalf of M1G Consulting Lid, (Company regis/raj Number 05406018) Abstract 2

Title of the Invention

I Background of the Invention 4

1.1 Global System for Mobile Communications GSM [1] 4 1.2 (ISM System architecture [3] 5 1.2.1 Mobile Station 1.2.2 Base Station Subsystem s 1.2.3 Network Subsystcm [3J 6 1.3 Multimedia Messaging Service (MMS) [4J 7 1.4 Al techniques in Pattern recognition isj 7 2 l)eseription of Related Art 9

3 Summary of the invention 11

4 Brief Description of the drawings 12

4.1 Video monitoring system (1) 12 4.2 Image Processing and decision support system (2) 12 4.3 GSM corn munication system (3) 15

Description of embodiments claims 17

References 18 Figures 19 Author (?tthis report Dr. Goran I lJezanoj, PhD l'his report is prepared under instruction from Mr. John I/endrickcon by the Author in accordance wi/h the agree,neni signed between them on 18/05/006 Report prepared by Dr. Goran / Bezanov PhD Jhr and on behalf of MiG Consulting Lid, Companv registration Number 05406018) References [1.1 [2] [4] uk.org. uk/rnms. htrn [5] n/Lcn.wi kipedi a.org/wi ki/Pattem recognition [6] Richard 0. Duda, Peter E. Hart, David G. Stork (2001) Pattern classification (2nd [7}Takashj Yoshiyama US Patent App!icatj US 2002/0057342A I {8J Rachel Furner UK Patent Application GB 22581 14A [9] Gonzales Ruben WIPO W0 01/06791 Al [10] Tanaka Koji Japan 2004-08029(4 edition), Wiley, New York, ISBN 0471056693.

[11] 13 1 78 [12] Jillp//www ln4rna.co. ilk/prod ucts/pock gsm. html [13] http://wwwndcom/ [14] Identify characteristics of video leeds http://wwwJldCom/apps/sciencejtl#4 Author 0/ this report Dr. (loran I Bezanov PhD This report is prepared under instruction from 1r. John Hendrickson by the Author in accordance with the agreement signed between them on /8/05/006 Report prepared by Dr. Goran 1 Bezanov PhD Jbr and on behalf of MIG Consulting Lid, (Company registration Number 05406018) Figures The following is the list of figures pertinent to this report: Figure 1 Layout of the Intelligent Mobile Security System 20 Figure 2 Main loop flowcht 21 Figure 3. Artificial Neural Netw(.)rk Flowchart 22 Figure 4 Additional Sensing Layout of the Intelligent Mobile Security System.... 23 Author ?! this report Dr. Goran I Bezanov I'hD This report is prepared under instruction from Mr. .John Hendrickson by the Author in accordance with the agreement signed between them on 18/05/006

Claims (6)

1. Report prepared by Dr. Goran I Bezanov PhD Jhr and on behalf of 411(1

   Consulting Lid, (Company registra/io Number 05406018) Descriphon of embodiments claims Innovation is claimed in the following: 1. A mobile remote monitoring security system comprising video and other sensor input, an interface to a personal computer which contains software programs with embedded intelligence to enable signal Processing and computer based decision making and a commuflicatjoii interface to transmit and receive information on the 3G GSM network.
2. 2. A mobile remote monitoring security system according to claim I, which is fitted with one or more web-cams 5Uppoing audio and video signal input, and a range of intruder detection Sensors
3. 3. A mobile remote monitoring security system according to claim I, which is fitted with fire and gas sensing equipment.
4. 4. A mobile remote monitoring security system according to claims I and 2 and 3, in which intelligent software algorithms are used to compare video images and to process sensor inputs and to make decisions as to whether or not a security threat exists.
5. 5. An intelligent mobile remote monitoring security system according to claims I and 2 and 3, which embodies computer based decision suppo to evaluate the security state of the system being monitored.
6. 6. A mobile remote monitoring security system according to any of the claims mentioned above, which is used to monitor activity nd determine the safety of occupants in a room where elderly or sickly occupants are situated.

   Author o/'Ihis report Dr. Goran I Bezanov PhD This report is prepared under instruction from Mr. John Hendrick.on by the Author in accordance with the agreemen, signed between them on 18/05/006 L1