CLOUD BASED INTELLIGENCE ARCHITECTURE - 2


There are 3 types of information stored in the platoons control terminal.(ground level).

  1. Raw data
  2. Meta data
  3. Beliefs.

This corresponds t o a model of each collectors(or platoon comprising of more than one collector) environment.My point is we can create an interface(Boot-level-mobile/fixed TOC?) which will contain all these information from all such collectors or platoons (to post messages for peer agent consumption at an indeterminate time.)so that information push-retreival/sharing can be achieved laterally and vertically with the Bn Staff intel./TOC.We can also allocate one storage system to each collectors terminal/platoons workstation so that its more easier to input/retrieve information from the common TOC.We can term these locations agents Notepad.All such Notepads are connected laterally and vertically with the Bn/Units Notebook.There will be a Notebook manager (Sr NCO) who will handle all storage and retrieval of information.There will be a system to ascertain the integrity of the data and will be under the purview of the truth maintenance system TMS.It will execute two functions:
1.Say we have a belief.Now beliefs are current--repeat current geospatial data.So if we have a belief that there is an observation at grid[x,y] at time (t0) and later at time (t1) there is no information for grid [x,y] then as the most recent value is of import, the easiest and simplest way to resolve conflicts between beliefs is to update the information on the notebook to the most recent belief and delete the other one.We can hire a software company and create more sophisticated conflict resolution algorithms.
2. Information must be accurate , timely , unique and specific to be translated to actionable intelligence.Keeping this in perspective 4 factors weigh in favor of this argument , viz: the age,proximity(to the sensor),uniqueness and operational relevance of the collected and stored information in the Bn Notebook.For operational success the unit needs to have rapid access to actionable intelligence.The TMS NCO must ensure that the most relevant timely information is cached for rapid access and if there is more data than that can be managed , (as it happens for imagery data..inundation like anything--we have hundreds and thousands of imagery data arriving from overhead sensors) then data that is old or less important is removed from the Bn notebook.Thus the second important function of the TMS is the efficient storage of information and that is accomplished by a properly directed caching and removal of information from the Bn Notebook.

For example platoon collectors collect raw information , that is still unprocessed and collected randomly , not yet indexed or categorized as per the exact prioritized intelligence requirements of the platoon commander.Yes the requirements have been set out in the mission tasking order but the collectors collected information in their own way and have pooled them together in draft form ,not yet assigning weights as per the priority rankings.

From this raw data ''sets'' are created or formed by judgement that contain information corresponding to a similar attribute. Now the sensor (collector) position , time and pose and other metadata gives meaning and context to this set of raw data.

Geospatial data such as targets , blue force locations,search areas are abstract information that are called beliefs. Platoon collectors (humans) , overhead sensors and pattern recognition software installed in surveillance platforms all contribute to the setup of beliefs repository. We can also have data fusion algorithms , fusing data from different sources.These are abstract facts about the current situation. Mission-level objectives are set down by the Commander and we can categorize this as a special class of beliefs.These beliefs drive the entire effort of the collection system , right from raw information to metadata providing context and meaning and the other beliefs.