MobileFusion, Inc.

All contestants:
MobileFusion, Inc.
We are a intelligent analytics company
Pittsburgh, Pennsylvania, United States United States
  • Overview
  • Team
Company description

One of the most difficult challenges decision makers facetoday is not access to data and information but being overwhelmed by it,whether that is a soldier searching for the enemy, a technician managing an oilrefinery or a driver optimizing the performance of his automobile. The variety,quality and number of available sources of data are ever increasing. Yet thechallenge remains: how to improve human real-time, real-world decision-making?  This challenge is $12 billion marketopportunity.

MobileFusion is an “intelligent analytics” company attacksthis opportunity through our core product, SmartFusionÔ, which is a modular softwareplatform designed to significantly improve decision making for anyone relyingon disparate data sources to make critical decisions in real-time. SmartFusionworks by fusing all available data – regardless of the number, complexity, typeand variety of sources – autonomously and continuously to generate simplerecommendations, alerts or both.

The core of SmartFusion is machine learning, whichautonomously processes very large amounts of disparate data in real-time muchthe way your brain continuously and seamlessly processes and fuses differentdata inputs (i.e., sight, sounds, smells, motion) to enable you to do everydaytasks like driving a car or cooking a meal. 

Business model

Due to the severity of the need and current funding levelsfor ISR systems our beachhead market is US Defense and Homeland Security.  As we gain traction we will move intothe adjacent US Commercial Security market.  Our sales and distribution strategy relies on partneringwith established companies in each target market.  We supplement those partnerships with direct touch to keyend customers to stimulate demand. For the US defense market this strategy involves partnering with primecontractors such as Boeing, Lockheed Martin, Textron, Stimulus Engineering, andAgile Business Concepts.  We havesupplemented this with additional individual contractors with direct ties intovarious Defense agencies and programs. 

MobileFusion’s primary revenue stream will come from sellingthe SmartFusion Intelligent Analytics platform packaged in one of two ways;

·     Standalone Software for end users – in thismodel will be implemented software with three revenue components 1) Softwarelicense, 2) implementation services and 3)

·     Embedded/OEM software to snap into largersystems – in this model software will be imbedded within a larger system.  Revenue will be from the followingthree components:  1) developerlicense, 2) implementation services, and 3) per product royalty

Based on this overall businessmodel the chart to the right outlines our financial projections through2012.  We expect our firststandalone software sales to commence in Q4-2009.

We will focus on the following three markets: 

Smart Structures

The number, variety and quality of sensors embedded in homes,factories, warehouses, and office buildings grows and will continue to grow atan accelerating rate. These sensors are mainly intended to monitor such thingsas power consumption and optimization, environmental conditions, security,building maintenance and general comfort.

However, there are severe constraints on adding more and moresensors as the “smart building” market expands. First, more sensors demand moreprofessionals monitoring those sensors (similar to the current GBOSS limitationmentioned earlier). This is an ever more expensive and, some case impossible,proposition.

Second, the limits of human attention in turn limits theeffectiveness of the monitoring systems to see complex, subtle, or infrequentpatterns or events (or all of the above) in sensor data. For example, someimportant patterns in say, power consumption and environmental conditions,change in ways so subtle they cannot be observed by human monitoring,regardless of the qualifications of the personnel or the sophistication,quality, variety and even number of sensors. –  that challenge grows exponentially as the number and varietyof sensor data grows. Further, completely new patterns are always emerging aswell that again cannot or won’t be seen by human monitoring.

Last, it is very difficult and, in some cases, impossible forhumans to integrate disparate sensor data in real-time and on a continuousbasis to perceive objects, patterns and events.  For example, changes in power consumption and environmentalconditions may suggest security issues and visa-a-versa.

What is needed is an intelligent software platform that can fusean unlimited number and variety of sensor streams on a continuous basis in real-timeto yield such alerts and recommendations as the following:

·   Powermanagement: manage power consumption in a building by predictive loadbalancing and optimization.

·   Reducecarbon footprint: optimize the usage of various power sources (electricity,solar and others) during the day.

·   Environmentalcomfort: optimize the environmental comfort variables by predictive devicecontrol, and optimized load balance. 

Preventive Maintenance on Equipment

Currently, the cost of maintenance is growing at an acceleratingrate. The current best preventative maintenance solutions require costlytechnicians to evaluate the state of the equipment on a calendar drivenschedule. This can be a limited, ineffective and expensive approach for variousreasons including inconsistency in the quality of maintenance professionals,infrequent inspection, constant adaptations and evolutions in the equipment toname a few. 

The SmartFusion platform can improve existing preventativemaintenance programs that are largely driven by scheduling on a calendar orsome other measure of hours of use, etc. It seems that if this is anappropriate application for Smart Fusion it would be to improve diagnostics andrecommend preventative maintenance when the equipment needs it instead of whenthe calendar recommends. 

What is needed is a software platform sitting between theequipment user and the equipment that is diagnosing problems and recommendingperformance enhancements on a continuous basis and real-time. 

What is a needed is an intelligent software platform thatprovides alerts and recommendations such as the following:

·   Predictingmeantime before equipment failure: Predictive software analyzes the controlI/O data and develops a data driven model to predict MTBF for variousequipment.

·   Optimizingequipment performance: Predictive software processes disparate and complexmonitoring output continuously to provide alerts to monitor, service, and evenreplace.

The system would function in the following manner:

We are pursuing such an opportunity through the US Navy tomaintain engines using our SmartFusion platform.

Multimedia Content Management

In 2008 486 exabytes (1018bytes) of digital content was produced and even that vast quantity is expectedto grow to more than 2,500 exabytes by the end of 2012 (IDC report 2009).  About thirty percent of this data isgenerated by media houses like Bertelsmann, BBC, NBS to name a few andthey and others are all challenged to organize, search and retrieval this data.

Search engines like Google and more recent applications such asVivisimo are largely text based search engines and do not yet offer verycomprehensive multimedia search because extracting meaningful metadata from multimedia data is a challenging cognitivetask.

A similar challenge exists in the intelligence community. Infact, the US intel community and NIST we are exploring the application ofSmartFusion to this problem of multimedia organization, search and retrieval.Specifically, we are working on a solution to solve the 1st W of theW4 problem – Who, Where, When, and Why. Byfusing video analysis, audioanalysis, NLP, and contextual information from other data repositories we cancreate virtual avatars (or who) of areal-live character on the fly andthen use this information to create metadata for organizing, search andretrieval.

Competitive advantage

MobileFusion's competitive advantage is in the core machine learning based technology. Specifically, SmartFusion’s architecture is distributableand hardware independent, which allows MobileFusion to quickly create specificimplementations and to meet a wide variety of customer challenges. TheSmartFusion software is a complete intelligent analytics platform with thefollowing core subsystems·     

·     Filtering Engine – It cleans,transforms and extracts relevant sensor data from the raw data coming from eachsensor using our proprietary Universal Data Port (UDP).  The UDP provides a common interfaceregardless of the sensor much like a simple USB port on a laptop.

·     Classification Engine – It selectsand applies multiple dynamic and static “classifier algorithms” to each datastream to identify objects of interest to the user. It does so by interpretingdistinctive characteristics of each data stream on a continuous basis.  The output of the Classification Engineis a real-time stream of “labels” (i.e., “that sounds like a dog barking,”) andprobabilities (i.e., “50% probability that this audio stream represents a dogbarking.”) from each sensor.

·     Consensus Engine – It processes the“labels” coming from the Classification Engine – in real-time and without lossof information – using dynamic weighting algorithms.  As the “labels” are processed at successive levels withinthe Consensus Engine, a single consolidated answer and probability for eachobject or event detected by the system is arrived at and presented to the user,again, on a continuous basis.

A major benefit of SmartFusion is that the accuracy anbreadth of application as well as the quality of information provided to theuser is only limited by the quality, number and variety of the available datasources.  The main benefit over anysingle data source systems (i.e., a security camera system in a bank) is thatSmartFusion is far more robust and fault-tolerant because it uses multiple datastreams to find objects, patterns & events even if it is occluded from oneor more of the systems sensors (i.e., a person hiding behind a wall).  In addition, SmartFusion monitors datastreams 7/24/365 providing constant monitoring reducing labor costs andincreasing system scalability.