Pentagram Research Centre Private Limited


An Indian Private Limited company incorporated at Hyderabad, India


Home | People | Departments | Products | Publications | Research | Journal | Conference | AllianceContact



  Pentagram Research Centre Private Limited organized pursuant to the 1956 Indian Companies Act, is executed effective as of date February 27, 1997, by the Directors (i) Chitra Govindarajan (G. Chitra) and (ii) Govindarajan Ethirajan (E.G. Rajan) with its corporate office in Hyderabad, India

Highway Crime Prediction / Prevention

  Threat Avoidance Control (TAC) Version 1.0


Basic Philosophy

Threats are inevitable since every entity is a threat to itself and to every other entity in a bounded or unbounded, spatial or functional neighborhood. Threats due to individuals or groups are intentional, whereas those due to natural elements and other living species are unintentional. Threat management and control is an ability naturally gifted to all living species. Living species intuitively understand prowling threats in a neighborhood, based on experience and retain their location or undertake relocation by getting into threat management and control mode.

Travelling Motorists Problem (TMP)

A motorist R intends to travel in a particular direction to reach his predetermined destination. He has to avoid certain zones which might pose security threat to him and at the same time he should not sway away from his destination. The security status of zones ahead of his current position is unknown to him. This is what we call as 'Travelling Motorists Problem (TMP)'.

Scope for solving TMP

The problem that arises in this context is whether it is possible to provide a motorist a software support system that predicts the security status of a zone at a particular instant of time based on knowledge about crimes committed by individuals and groups over a period of time and other collateral information and suggest him a safe route which we call as ‘escape route’.


Lattice Grid Model for solving TMP

The basic model that is used in developing the system is described below. Let us assume that R→L1 is the intended travel direction of a motorist. The first threat possibility is due to a single security threat posed to the motorist from the direction L0R. This threat position is pictorially represented in figure given below. The presence of a 1 in a cell indicates the presence of a threat in that zone.

Now the motorist R can change his direction of motion to any one of the following ones: RL1, RL2, RL3, RL4. They are called ‘escape routes’ R can also adjust his speed depending on the threat level at zone L0. The motorist would confront 32 types of threat positions at any location at any instant of time. These 32 threat positions are shown below.


Legend: 0 → Threat-free zone; 0 → Threat prone zone; 1→ Threat zone; 0 → Unknown zone


At every moment and location, the motorist has to decide his escape route based on the a priori knowledge about the current threat position. Unfortunately the motorist may not  know the security issues associated with the spatial neighborhood of his current location and moment.This integrated product which is an expert system dynamically estimates security status in terms of 'Threat Indicator' to the motorist and hence it plays a very significant role ensuring safe and threat free travel to all those motorists who use this facility.

  Background Technology Preliminaries

The basic model proposed here shows two adjacent zones L3 and L4 and three zones L0, L1 and L2 ahead of the present location of a motorist. Each zone could be assumed to be of one square kilometer area (this can be changed by the user using his hand held device). Basic assumption before getting into further discussion is that security threat is all pervading and it is a time variant phenomenon. Threat-free zone is a momentary status and so is the permanent threat zone. The motorist has to decide his safe route (called escape route here) based on a priori knowledge about the surrounding zones which he is unlikely to have and predict current situation. Nor he has time to carry out huge calculations before getting into next zone. In this context, this software support system would take up this responsibility of predicting what could happen to him if he gets into a zone which may be safe or hazardous for him. The system evaluates the threat index for every surrounding zone in a chosen neighborhood and suggests the best safe zone with least threat index. The calculation is carried out dynamically at every chosen time and location of the motorist using the time varying crime statistics and other required collateral information associated with chosen zones for the past chosen periods of time. All these amount to saying that the calculation or prediction of a threat index of a zone is a very complex operation which involves machine learning techniques. On the need to know basis of the motorist, the system would choose a binary hypothesis testing or an M-ary hypothesis testing to predict threat index of a zone. These two hypothesis testing are outlined below very briefly.

Binary and M-ary Hypothesis Testing

Binary observation space is divided into H1 and H0. L(y) is the predicted threat index and it is either a 1 or a 0. M-ary observation space is divided into H0, H1, H2,...., HM-2, HM-1. L(yj) is the predicted threat index which ranges from 0 to M-1.


In the case of binary hypothesis testing, threat-free zone is suggested to the motorist for the next move. In the case of M-ary hypothesis testing, the zone with the minimum threat index is suggested for the next move. Thus a safe route is shown to a motorist continuously by the system present in the hand held device of the motorist. A stringent reliability testing of this software support system is under way now. However, overall reliability depends on the genuineness of the crime statistics provided periodically by the security and other agencies.


Digital Food Initiative (DFI)►  ◄Digital Health Initiative (DHI)►  ◄Digital Security Initiative (DSI)