GAJA — Proposal | TRiDE Innovative Technologies

TRIDE

CONTENTS

01Deployment Gap Analysis — Where GAJA Fits

07Layer 3 — Thermal AI Confirmation

02Problem Statement & Existing System Limits

08KAVACH Integration

03Complete Architecture Diagram

09Deployment Plan

04Working Principle — DAS

10Permission & Regulatory Profile

05System Architecture — Three Layers

Annex ABudget — Pilot Phase Detailed BoQ

06Layer 1 — Orbital Intelligence & CRS

Annex BNational Scale Economics & References

01 · DEPLOYMENT GAP ANALYSIS

Where GAJA Fits — Uncovered Corridors

Indian Railways has deployed or sanctioned DAS-IDS across 1,158 route km. GAJA targets the corridors not yet covered — and adds a predictive intelligence layer no existing system provides.

Indian Railways has made significant progress on elephant safety. As of early 2026, the AI-enabled DAS-based Intrusion Detection System (IDS) is active on 141 route km under NFR, with works sanctioned for an additional 1,017 km across 8 railway zones. Madukkarai (Tamil Nadu) has deployed a separate AI thermal camera system since February 2024 with zero elephant deaths in its first year. EleSense thermal sensors are deploying in West Bengal's Chapramari Wildlife Sanctuary.

However, this still leaves significant geographic and technical gaps that GAJA is specifically designed to address. The table below maps the current coverage status and the specific corridors where GAJA can be proposed without competing with existing deployments.

Current Coverage Status — Zone-wise

Railway Zone	Sanctioned (RKm)	Active (RKm)	Status	States Covered	GAJA Opportunity
NFR (Northeast Frontier)	403	141	PARTIALLY ACTIVE	Assam, Meghalaya, NE states	Predictive CRS layer still absent. 262 RKm sanctioned but not yet active.
ECOR (East Coast Railway)	369	~349	INSTALLATION IN PROGRESS	Odisha, Andhra Pradesh	DAS only — no predictive satellite layer. No thermal confirmation cameras.
SR (Southern Railway)	56	~56	INSTALLATION IN PROGRESS	Tamil Nadu, Kerala (Madukkarai separate)	Limited. Madukkarai AI system covers only 7km. Remaining SR corridors uncovered.
NER (Northeast Railway)	99	~36	INSTALLATION IN PROGRESS	UP, Bihar forest zones	63 RKm still pending. No predictive layer anywhere.
NR (Northern Railway)	52	0	SANCTIONED ONLY	Uttarakhand (Rajaji–Corbett corridor)	Nothing active. No AI deployed. High priority target for GAJA.
SER (Southeast Railway)	55	0	SANCTIONED ONLY	West Bengal, Jharkhand	Nothing active. EleSense in Chapramari (West Bengal) is separate private initiative.
SWR/WR (South Western / Western)	115	0	SANCTIONED ONLY	Karnataka, Goa, parts of Maharashtra	Nothing active. Karnataka forest corridors entirely unprotected currently.
ECR (East Central Railway)	20	0	SANCTIONED ONLY	Jharkhand, Bihar	Only 20 RKm sanctioned despite significant elephant presence in Jharkhand.
Chhattisgarh corridors	0	0	NOT COVERED	Chhattisgarh	Completely uncovered. Growing elephant population. No zone has sanctioned IDS here.
Remaining 77 − 15 priority corridors	~1,600+	0	NOT COVERED	Various states	MoEFCC identified 77 priority stretches — current IDS sanctions cover roughly 15–20 of these.

CRITICAL GAP — WHAT NO EXISTING SYSTEM PROVIDES

Even in zones where IDS-DAS is active or being installed, every existing system is reactive only — alerting 1–5km in advance when the elephant is already near the track. No deployed system anywhere in India provides: (1) predictive risk scoring hours before a crossing event, (2) satellite-based corridor intelligence, (3) KAVACH speed enforcement integration, or (4) GPS collar-to-CRS data fusion. GAJA's predictive intelligence layer is complementary to, not competing with, the existing DAS-IDS deployments — it can be layered on top of or alongside any active zone.

GAJA's Primary Deployment Targets

🎯 Priority 1 — Completely Uncovered Zones

Chhattisgarh (no IDS sanctioned despite ~60 elephants and growing), remaining NR Uttarakhand (Rajaji–Corbett corridor, 52 RKm sanctioned but nothing active), SWR Karnataka corridors (115 RKm sanctioned, nothing deployed). These are the highest-priority first targets for GAJA.

🎯 Priority 2 — Sanctioned but Not Yet Deployed

SER (West Bengal/Jharkhand), ECR (Jharkhand/Bihar), NER remaining 63 RKm. Works sanctioned means DAS hardware is coming — GAJA can be proposed as the predictive intelligence and thermal confirmation layer to be integrated before those installations are commissioned.

🎯 Priority 3 — Enhancement of Active Zones

Active NFR zones and ECOR installations have DAS but no predictive CRS, no satellite layer, no KAVACH enforcement, and no thermal visual confirmation. GAJA's Layer 1 and Layer 3 can be proposed as enhancements to the existing DAS backbone — different procurement, different budget head, different system owner (could be MoEFCC rather than Railways).

02 · PROBLEM STATEMENT

The Crisis & Limits of Existing Systems

Existing systems alert 1–5km in advance — a significant improvement. But the physics of braking and the biology of elephant behaviour demand more.

186+

Elephants killed since 1987

Deaths 2019–2024 alone

1–5km

Existing DAS-IDS alert range

600–800m

Full stop distance at 60 km/h

30 km/h

Mandated night speed in forest zones

Corridors with predictive AI layer

What Existing Systems Do Well

The NFR DAS-IDS system has achieved a remarkable result: over 160 elephant lives protected in 2025 across 62.7 km of active sections. The Madukkarai AI camera system recorded zero elephant deaths in its first year of operation (Feb 2024 – Feb 2025) with 5,011 alerts generated and 2,500 safe crossings facilitated. These are genuine achievements that must be acknowledged.

Existing systems alert loco pilots and control rooms when an elephant is detected within approximately 1–5 km of the train's position — providing 1–5 minutes at 60 km/h to slow the train. This is far better than the 2–3 seconds available with no system. At 30 km/h (the mandated night speed), 1km of warning gives 2 full minutes — enough time to stop safely.

What Still Needs Solving

No Predictive Intelligence Anywhere

Every existing system — DAS-IDS, Madukkarai cameras, EleSense — is reactive. None uses historical crossing patterns, satellite data, GPS collar positions, or environmental variables to predict when and where a crossing will occur hours in advance. Elephants follow fixed ancestral corridors and are 60–80% predictable — this intelligence is entirely unexploited.

Speed Advisory Non-Compliance

The 30 km/h night speed mandate in forest zones is driver-discretionary. No automated enforcement exists. Multiple post-collision investigations found trains above the advisory speed. KAVACH exists but has no wildlife-interface module — no system has integrated DAS alerts with automated speed enforcement through KAVACH.

60+ Uncovered Corridors

MoEFCC identified 77 priority stretches. Current IDS sanctions cover roughly 15–20 of these with active systems. Chhattisgarh has zero coverage despite a growing elephant population. Karnataka, Uttarakhand, large portions of Jharkhand and Bihar remain completely unprotected.

No Multi-Stakeholder Data Platform

Existing systems alert the loco pilot and occasionally the station master. No system currently feeds live data simultaneously to OCC, Forest Department, MoEFCC Project Elephant, and the Railway Safety Commissioner in a unified dashboard. Inter-agency data sharing happens after collisions, not before.

03 · ARCHITECTURE DIAGRAM

Complete System Architecture

All physical components, data flows, and system connections from forest approach to loco cab — in one view

The diagram shows: top zone — 6 orbital data sources feeding the CRS engine; middle zone — predictive AI, TSR advisory, DAS interrogator, AI classifier, OCC dashboard, and all 5 alert output channels; bottom zone — physical field deployment showing elephant approaching from left forest, two OHE masts with thermal cameras, railway track with sleepers and rails, OFC fiber cable along track, early warning OFC deeper in RoW, underground conduit, station building with DAS rack, locomotive with cab alarm active, and RoW boundary lines.

04 · WORKING PRINCIPLE

Distributed Acoustic Sensing — The Fiber as a Sensor

Converting Railway's existing OFC into a continuous 100km vibration sensor from inside the station building

Laser Pulse Injection

A coherent laser pulse is fired into a single core of the existing Railway OFC cable from the DAS interrogator unit installed inside the nearest station equipment room. No trackside installation. No new cable on corridors where Railway OFC already exists.

Rayleigh Backscattering

Microscopic glass imperfections scatter light back toward the source. Time-of-return determines distance to ±3 metres along 100km of cable. One interrogator unit covers the entire monitored section — the fibre is the sensor, not a conduit for sensor data.

Vibration Detection & Signature

A 5-tonne elephant creates a highly distinctive low-frequency vibration: 4–12 Hz dominant, regular 0.8-second footfall cadence, heavy amplitude. This signature is clearly distinct from trains (broadband, very high energy, continuous), cattle (lighter mass, irregular), humans (high frequency, low energy), and wind (random broadband).

Reference Baseline & Continuous Comparison

On deployment, a reference acoustic fingerprint captures the corridor under idle conditions. Train signatures are catalogued and permanently filtered. All real-time traces compare against this baseline — deviations exceeding threshold trigger the classifier. The system self-calibrates seasonally as environmental conditions change.

AI Classification

ML model (Random Forest + LSTM temporal classifier) categorises each event: Elephant Herd / Single Elephant / Train (filtered) / Human-Cattle / Digging-Encroachment / Fiber Cut. Model trained on NFR pilot acoustic data and fine-tuned per corridor. False positive target: under 2 per 100km per 24 hours.

Classification latency: <200ms from vibration detection

Alert Generation — Sub-1.5 Second End-to-End

Alert packet includes: GPS location ±3m, confidence %, herd size estimate, nearest train distance, time since detection, CRS context. Transmitted via 4G/LTE to OCC, loco cab, patroller app, Forest Dept. Same fiber simultaneously tracks all trains on the corridor — enabling targeted alerts only to approaching trains.

Total end-to-end alert latency: <1.5 seconds

HOW GAJA DIFFERS FROM EXISTING DAS-IDS (1–5KM vs 2–6 HOURS)

The existing NFR DAS-IDS alerts loco pilots and station masters when an elephant is detected near the track — typically within 1–5km of the approaching train. This is valuable and has saved 160+ elephant lives in 2025. GAJA keeps this DAS proximity layer and adds the predictive intelligence layer that no existing system has — using satellite data and pattern AI to issue speed restrictions 2–6 hours before the elephant approaches the track. These are complementary, not competing systems.

05 · SYSTEM ARCHITECTURE

Three-Layer Architecture

Each layer compensates for the failure modes of the others — at 30 km/h, stopping distance equals thermal camera range

Orbital Intelligence — Prediction 2–6 Hours Ahead

Satellite stack + GPS collars + pattern AI + environmental data → Corridor Risk Score

Computes a 0–100 Corridor Risk Score (CRS) per 10km segment every 90 minutes. When CRS exceeds 70, a Temporary Speed Restriction (TSR) is auto-issued to the OCC dispatcher — proactively, hours before the elephant reaches the track. At 30 km/h, the train's stopping distance drops to 150m, within thermal camera range.

CRS Score	Status	Automated Action	Train Speed
0–40	● LOW	No restriction. Standard monitoring.	Line speed
40–70	● ELEVATED	Advisory to OCC. KAVACH cab display alert.	Line speed with caution
70–90	● HIGH	TSR auto-issued. 30 km/h. Patroller SMS. Forest Dept notified.	Max 30 km/h
90–100	● CRITICAL	TSR + possible block closure. DRM notified. All patrollers deployed.	Max 15 km/h or halt

DAS Fiber Sensing — Real-Time Ground Detection

Existing Railway OFC → 100km continuous sensor → elephant detected 30–40m from cable

Two-cable configuration: an early warning OFC (~100ft from track, within RoW) for 60–200 second advance detection, and the standard proximity OFC at track edge for high-confidence braking alert. Same fiber simultaneously tracks all train positions on the corridor — giving OCC a unified live map of trains and wildlife.

Thermal AI Confirmation — 150m Visual Verification

DAS-triggered LWIR cameras on existing OHE masts — no new poles, no forest land

Activated only when DAS fires — not always-on. Cameras wake within 200ms of DAS alert. YOLO-based edge AI confirms elephant species, count, and trajectory within 200ms. At 30 km/h (the proactive restricted speed), 150m gives 18 seconds — full braking margin. In-cab alarm fires only on thermal + DAS dual confirmation, minimising false positives.

Real-Time Alert Chain

〰️

0 ms

DAS detects footfall. Location ±3m. Classifier triggered.

🔴

+200ms

Nearest OHE thermal cameras activate. YOLO confirms elephant.

📡

+600ms

Alert packet: location, confidence, herd size, train distance → OCC via 4G.

🚂

+900ms

Loco cab alarm. KAVACH geofence. Patroller SMS. Forest Dept. Station Master.

✅

<1.5s

Complete. At 30 km/h: 150m stop distance = thermal camera range.

06 · LAYER 1 — ORBITAL INTELLIGENCE

Satellite Stack & Predictive AI

Multi-provider satellites + GPS collar telemetry + 35 years of collision data → Corridor Risk Score every 90 minutes

Elephants use fixed ancestral corridors and cross railway tracks at consistent locations, seasonally, at night. Their movement is 60–80% predictable with sufficient historical data. No existing system exploits this. GAJA's Layer 1 is the first application of satellite intelligence and pattern AI to Indian Railways elephant safety.

Source	Type	Coverage	GAJA Use
Capella Space / ICEYE	SAR X-Band radar	Every 1–3 hrs, any weather, night	Cloud-penetrating detection of large animal clusters 2–5km from track. Active during monsoon when optical satellites are blind.
Satellite Vu	Thermal IR 3.5m	4–6 passes/day	Elephant herds show as heat clusters. First commercial thermal IR satellite capable of detecting large mammals from orbit.
Planet SkySat	Optical 50cm	4–6 passes/day	Visual corridor mapping, NDVI vegetation monitoring, habitat change detection in approach zones.
ISRO Cartosat-3	Optical 28cm	Daily	High-resolution baseline mapping. Available at reduced cost for Government Railways projects through NRSC/ISRO.
GPS Collar Telemetry	Real-time GPS	Every 4 hours	Direct API integration with WWF India, WII, and State Forest Dept collar databases. Herd proximity to track is the strongest single CRS predictor.
Historical Collision DB	Event records	1987–2024	187 collision events with GPS coordinates, time, season, weather. Bayesian prior for crossing probability at each 10km segment.
Environmental Variables	NDVI, rainfall, moon, crops	Daily / real-time	Moon phase, post-monsoon harvest calendar, rainfall anomalies, and NDVI (via free Sentinel-2) are strong secondary CRS predictors.

07 · LAYER 3 — THERMAL CONFIRMATION

Thermal AI Cameras on OHE Masts

DAS-triggered, OHE-mounted, YOLO-classified — 150m detection, zero forest land, zero always-on power draw

OHE Mast Mounting

OHE masts exist every 60–80m along electrified track — all Railway-owned, all within RoW. Thermal cameras bolt onto existing mast arms via a custom stainless clamp bracket. Power tapped from 240V AC OHE mast supply. No new poles, no foundations, no forest land. Requires only OHE dept NOC.

Camera Specification

FLIR Boson+ 640×512 or equivalent, 8–14µm LWIR, 50mm lens, IP67 housing, −40°C to +80°C operating range. At night, elephant body temperature (~35°C) against forest background (~22°C) produces a high-contrast signature detectable at 120–180m with 98%+ confidence.

Edge AI — YOLO on Jetson

Each camera pairs with a Jetson Orin Nano edge node running fine-tuned YOLOv9 on Indian elephant thermal imagery. Classifies species, estimates count, tracks trajectory. Inference: under 50ms at 15fps. Results + 30-second thermal clip transmitted to OCC for audit trail and ML retraining.

DAS-Triggered Operation

Camera sleeps at ≤0.8W. DAS alert wakes nearest 2–3 cameras within 200ms. If no elephant confirmed within 30 seconds, alert cancelled and logged as false positive — feeding monthly model retraining. This design reduces false positives by 90% vs always-on camera-only systems like Madukkarai.

08 · KAVACH INTEGRATION

Automated Speed Enforcement

Closing the advisory compliance gap — trains physically cannot exceed geofenced speed in high-risk zones

KAVACH (Train Collision Avoidance System) is being deployed nationwide by Indian Railways. GAJA integrates with KAVACH as an external hazard data source — the only wildlife safety system to do so. This closes the single most critical gap in the current approach: speed advisory non-compliance.

Dynamic Geofence Speed Zones

GAJA pushes geofenced speed limits to KAVACH's track database when CRS >70. Trains physically cannot exceed 30 km/h in the zone — KAVACH enforces via automatic brake application if the driver fails to comply. No discretion, no fatigue, no override.

Automatic Brake on Critical Alert

When DAS + thermal confirm a CRITICAL event (CRS >90), GAJA can send a braking command to KAVACH for trains within 2km. KAVACH applies brakes automatically — same mechanism as signal-at-danger violation prevention. Driver receives voice warning 3 seconds before automatic brake.

Cab Display Integration

On KAVACH-fitted locos, GAJA alerts appear directly on the existing cab display screen. No additional hardware. Alert shows: corridor segment, distance to elephant zone, recommended speed, and "ELEPHANT ZONE AHEAD" voice warning in the local language.

Phase 2 Deployment

KAVACH integration requires RDSO approval — typically 6–12 months. GAJA is fully operational without KAVACH. Phase 0 and Phase 1 use radio-based driver alert. KAVACH enforcement adds automated compliance to an already-functional alerting system.

09 · DEPLOYMENT PLAN

Phased Rollout — Uncovered Corridors First

Starting where no AI system currently exists — Chhattisgarh, Uttarakhand, Karnataka, remaining SER/ECR gaps

Phase	Target Corridors	Components	Duration	Permissions
Phase 0 Data Baseline	One uncovered corridor: Chhattisgarh (Raipur/Bilaspur division) or NR Uttarakhand (Rajaji–Corbett). Minimum 50km section with existing OFC.	DAS interrogator · OCC dashboard · Satellite subscription · CRS model training on local collision history	60–90 days	Railway Telecom NOC only
Phase 1 Pilot Live	Same corridor — full three-layer system. Thermal cameras at 5 highest-risk crossing points on OHE masts. Patroller app with Forest Dept.	Phase 0 + Thermal cameras + Edge nodes + Driver radio alert + Patroller app + Forest Dept data MoU	3–4 months	Telecom NOC + OHE dept NOC
Phase 2 Expansion	3 additional uncovered zones: SWR Karnataka, ECR Jharkhand, SER remaining West Bengal. GPS collar API with State Forest Depts. KAVACH integration spec begins.	Phase 1 × 3 + GPS collar MoU + KAVACH API development + MoEFCC Project Elephant data feed	6 months	Forest Dept MoU (data only)
Phase 3 Platform	All remaining uncovered corridors from the 77 MoEFCC priority stretches. KAVACH enforcement live. National CRS dashboard for Project Elephant and Railway Safety Commissioner.	Full platform + KAVACH integration + national dashboard + RDSO certification + MoEFCC reporting	18–24 months	RDSO approval for KAVACH

10 · REGULATORY PROFILE

Permission Matrix

Designed to avoid the two historic blockers: forest land and track infrastructure approval

Component	Forest Dept	Railway Track Approval	RDSO	New Land	Timeline
DAS on existing OFC	NOT REQUIRED	Telecom dept NOC	NOT REQUIRED	Zero	30–45 days
Satellite subscription	NOT REQUIRED	NOT REQUIRED	NOT REQUIRED	Zero	2–4 weeks
Thermal cameras on OHE masts	NOT REQUIRED	OHE dept NOC	Safety assessment	Within RoW	30–45 days
OCC Dashboard	NOT REQUIRED	NOT REQUIRED	NOT REQUIRED	Zero	2–3 weeks
GPS collar data API	MoU (data sharing only)	NOT REQUIRED	NOT REQUIRED	Zero	2–4 months
KAVACH integration	NOT REQUIRED	NOT REQUIRED	Required (Phase 2+)	Zero	6–12 months

ANNEXURE A

Detailed Budget — Phase 0 + Phase 1 Pilot

50km uncovered corridor · Indicative estimates · Subject to site survey, GeM quotation, and Railways procurement norms

DISCLAIMER

All figures are indicative estimates for proposal planning. Final costs subject to detailed site survey, vendor quotation via GeM/tender, and applicable Railway procurement norms. GST as applicable (18% on equipment, 12% on services) not included. USD-denominated satellite subscriptions subject to exchange rate variation.

Item Description	Qty	Unit	Unit Rate (₹)	Amount (₹)	Notes
A. DAS INFRASTRUCTURE
DAS Interrogator Unit — coherent OTDR, 100km range, rack-mounted 2U, SMF-28 compatible, includes laser source and signal processing card	1	Unit	45,00,000	45,00,000	Sensonic/AP Sensing/equiv
DAS AI Server — 1U rack, GPU-accelerated (NVIDIA RTX A2000), 64GB RAM, 4TB NVMe, redundant PSU, 3-yr warranty	1	Unit	12,00,000	12,00,000	Dell/HP enterprise
OFC splicing and termination at station — splice into existing Railway OFC for DAS tap, termination box, patch panel, OTDR testing, documentation	1	LS	2,50,000	2,50,000	Railway-approved contractor
Network switch (managed 24-port GbE), UPS (2 KVA online), 42U rack enclosure, PDU, cable management	1	LS	2,20,000	2,20,000	APC/Schneider
4G/LTE modem (primary) + satellite backup link for OCC uplink — ensures alert delivery even in network outage	1	LS	1,80,000	1,80,000	Backup critical for remote corridors
OFC health audit — 50km OTDR testing of existing Railway OFC, documentation of fiber condition, identification of high-loss sections	1	LS	4,50,000	4,50,000	Required pre-DAS installation
Sub-total A — DAS Infrastructure				68,00,000
B. THERMAL CAMERA SYSTEM (LAYER 3) — 5 cameras at identified crossing points
LWIR Thermal Camera — FLIR Boson+ 640×512, 50mm lens, 8–14µm waveband, IP67, −40°C to +80°C, with calibration certificate	5	Nos	3,80,000	19,00,000	One per crossing point
OHE mast clamp bracket — custom stainless steel, adjustable azimuth ±30°, hot-dip galvanised, weatherproof junction box	5	Nos	58,000	2,90,000	Railway-approved fabrication
Edge compute node — NVIDIA Jetson Orin Nano 8GB, IP65 enclosure, DIN rail, −20°C to +70°C, fanless, YOLOv9 pre-loaded	5	Nos	95,000	4,75,000	One per camera location
Power supply unit — 12V DC regulated 30W, DIN rail, surge protection, isolation transformer for OHE mast 240V AC tap	5	Nos	30,000	1,50,000	Including MCB and earthing
Armoured power + data cable (15m per location), conduit, weatherproof connectors	75	Mtr	1,200	90,000	15m × 5 locations
Camera installation, alignment, commissioning, acceptance testing per location	5	Nos	38,000	1,90,000	Includes OHE dept safety permit
Sub-total B — Thermal Camera System				30,95,000
C. SATELLITE INTELLIGENCE SUBSCRIPTION (ANNUAL — LAYER 1)
SAR satellite tasking — Capella/ICEYE, 50km AOI, 2 passes/night minimum, 25–50cm resolution, API delivery <15min	1	Year	20,00,000	20,00,000	USD-denominated
Thermal IR satellite — Satellite Vu, 50km AOI, 4–6 passes/day, 3.5m resolution, elephant heat detection API	1	Year	14,00,000	14,00,000	Indicative — emerging provider
Optical satellite — Planet SkySat, 50km AOI, NDVI analytics, change detection, corridor mapping baseline	1	Year	9,00,000	9,00,000	Planet Gov pricing applicable
ISRO Cartosat-3 — quarterly tasking, 50km AOI, 28cm resolution, baseline corridor and infrastructure mapping	1	Year	2,00,000	2,00,000	Near-zero cost via NRSC for Govt projects
Sub-total C — Satellite Subscriptions (Annual)				45,00,000
D. SOFTWARE PLATFORM & AI DEVELOPMENT
GAJA Platform licence — CRS engine, OCC dashboard (web + mobile), DAS API, alert management, event log, annual licence with updates and support	1	Year	15,00,000	15,00,000	TRiDE Innovative Technologies
DAS elephant model — initial training + corridor fine-tuning, deployment, validation, 30-day acceptance testing, documentation	1	LS	8,00,000	8,00,000	One-time development
Thermal YOLO model — fine-tuning on Indian elephant thermal dataset, edge deployment on 5 Jetson nodes, validation	1	LS	4,50,000	4,50,000	One-time per config
CRS historical data ingestion — geocoding of local collision records, corridor shapefile integration, initial model training	1	LS	3,00,000	3,00,000	One-time data preparation
Patroller mobile app — Android, offline-capable, bilingual, GPS alert, herd reporting, false positive marking	1	LS	3,50,000	3,50,000	For Forest Dept team
OCC hardware — 55" 4K display, OCC workstation (i7, 32GB, GPU), installation and configuration	1	LS	2,80,000	2,80,000	At nearest divisional OCC
Sub-total D — Software & AI				36,80,000
E. CIVIL, INSTALLATION & COMMISSIONING
Station equipment room fitout — cable routing, conduit, earthing, AC (1.5T split), fire extinguisher, labelling	1	LS	3,50,000	3,50,000	At DAS station building
System integration, end-to-end testing, 30-day acceptance trial — DAS to OCC alert chain, thermal trigger, CRS validation, false positive rate measurement	1	LS	5,00,000	5,00,000	30-day acceptance period
Staff training — OCC operators (2 days), patrollers (1 day), Railway S&T maintenance (1 day), SOPs preparation	1	LS	2,00,000	2,00,000	Including bilingual training materials
TRiDE project management and resident engineer — Phase 0 + Phase 1 (approximately 5 months)	5	Months	1,20,000	6,00,000	Resident engineer + travel
Sub-total E — Civil & Commissioning				16,50,000
TOTAL CAPEX (One-time) — A + B + D + E				1,52,25,000
TOTAL ANNUAL OPEX — C (Satellite) + Platform Licence				60,00,000
TOTAL YEAR 1 (CAPEX + OPEX)				₹ 2,12,25,000	~₹2.12 Crore

₹1.52 Cr

One-time capital expenditure

DAS + thermal cameras + software + civil works

₹60 L

Annual operating expenditure

Satellite stack + GAJA platform licence

₹2.12 Cr

Total Year 1 investment

Complete pilot — 50km uncovered corridor · all three layers live

A single elephant-train collision costs ₹2–8 Cr in locomotive damage, investigation, delay costs, and compensation. GAJA Year 1 at ₹2.12 Cr costs less than one avoided collision.

ANNEXURE B

National Scale Economics & References

Unit economics · Phase-wise rollout · Source documents and research papers

Per-Corridor Unit Economics (100km steady-state)

Component	Qty / 100km	Unit Cost (₹)	Total (₹)	Notes
DAS Interrogator (100km range)	1	45,00,000	45,00,000	One unit covers full 100km
DAS Server + AI Processing	1	12,00,000	12,00,000	Shared per section
Thermal Cameras (5km intervals)	20	5,10,000	1,02,00,000	Camera + edge node + mount + cabling
GAJA Platform (annual)	1/yr	12,00,000	12,00,000	Volume discounted from pilot rate
Satellite Intelligence (annual, 100km AOI)	1/yr	60,00,000	60,00,000	SAR + Thermal IR + Optical stack
Installation + commissioning	1	20,00,000	20,00,000	OFC audit + camera install + testing
TOTAL CAPEX per 100km corridor			1,91,00,000	One-time
TOTAL OPEX per 100km / year			72,00,000	Satellite + platform + maintenance

Phase-wise National Rollout

Phase	Corridors	km	Capex (₹ Cr)	Annual Opex (₹ Cr)	Yr 1 Total (₹ Cr)
Phase 0+1 — Pilot (1 uncovered corridor, 50km)	1	50	1.52	0.60	2.12
Phase 2 — 4 uncovered corridors (~400km)	4	400	14.0	5.0	19.0
Phase 3 — 20 uncovered/partial corridors	20	2,000	62.0	22.0	84.0
Phase 4 — All remaining uncovered stretches	55+	~5,500	165.0	58.0	223.0

References — Government Sources

R1Ministry of Railways, Rajya Sabha Written Reply — AI-enabled IDS deployment status, sanctioned corridors (1,158 RKm, ₹208 Cr). PIB Release, February 2026.
https://www.pib.gov.in/PressReleasePage.aspx?PRID=2223196
R2Ministry of Railways, Lok Sabha Written Reply — IDS-DAS deployment, ECOR/SR/NER installation progress. PIB Release.
https://www.pib.gov.in/PressReleasePage.aspx?PRID=2113870
R3MoEFCC — 2-Day National Workshop on Elephant-Train Collisions, March 12–13, 2026. ANI/PIB Coverage.
https://aninews.in/news/national/general-news/environment-ministry-organizes-2-day-workshop-
R4NFR CPRO — "160+ Elephant Lives Protected in 2025 Using DAS-IDS." Northeast Frontier Railway Official Statement, December 2025.
https://www.newkerala.com/news/o/nfrs-tech-driven-measures-save-over-160-elephants-ensure-519
R5MoEFCC Project Elephant — "Elephant Corridors of India 2023." Ministry of Environment Report identifying 88 corridors across India.
https://moef.gov.in/uploads/2023/11/PE-Elephant-Corridor-of-India-2023.pdf

References — Madukkarai AI System

R6Supriya Sahu, Tamil Nadu APCEF — "Zero Elephant Accidents, 5,011 AI Alerts, 2,500 Safe Crossings." Madukkarai AI system results, February 2025. Down to Earth.
https://www.downtoearth.org.in/wildlife-biodiversity/ai-enabled-system-successful
R7India AI — "Tamil Nadu Officially Launches AI to Prevent Elephant Death on Railway Tracks." System specifications: 12 towers, 500m spacing, 150m coverage each side.
https://indiaai.gov.in/article/tamil-nadu-officially-launches-ai-to-prevent-elephant-death

References — Elephant Movement Research

R8Roy, M., Baskaran, N., & Sukumar, R. (2009). "The Death of Jumbos on Railway Lines in Northern West Bengal." Gajah, 31, 36–39. Historical analysis of elephant-train collisions in the Siliguri–Alipurduar corridor.
R9PMC / MDPI Sensors, 2025 — "An Analysis of Railway Activity Using Distributed Optical Fiber Acoustic Sensing." FOSINA study: 12km DAS deployment detecting wild boar, deer, humans, and aircraft. Demonstrates DAS animal classification via ML.
https://pmc.ncbi.nlm.nih.gov/articles/PMC12252076/
R10PMC / 2025 — "Memory-Based Navigation in Elephants: Implications for Survival Strategies and Conservation." Confirms elephants use fixed ancestral corridors, time railway crossings at night in Rajaji NP, and alter routes based on learned spatial memory — supporting the predictive intelligence approach.
https://pmc.ncbi.nlm.nih.gov/articles/PMC12030947/
R11WWF India — "GPS Collars Show Elephant Movements." GPS collar telemetry of 5 collared elephants in Assam demonstrating seasonal migration, harvest-season crop-raiding, and cross-border movement patterns.
https://www.worldwildlife.org/news/stories/how-elephants-navigate-indias-changing-landscape/
R12Wildlife SOS — "Elephant Migration: The Journey of Gentle Giants." Documents 18 inter-state migratory routes between Tamil Nadu and Kerala, seasonal patterns, and Chhattisgarh corridor — relevant to GAJA target zones.
https://news.wildlifesos.org/elephant-migration-the-journey-of-gentle-giants/

References — DAS Technology

R13Sensonic — "Distributed Acoustic Sensing for Railways Explained." Covers: 100km coverage per unit, animal intrusion detection, elephant-specific applications, East Coast Railway India project.
https://www.sensonic.com/en/blog/distributed-acoustic-sensing-for-railways-explained--3240/
R14AP Sensing — "Distributed Acoustic Sensing (DAS)." Technical reference: coherent OTDR principles, classification algorithms, spatial resolution, range specifications.
https://www.apsensing.com/en/technology-and-products/distributed-acoustic-sensing
R15Bitcomm Technologies — "Vibrations along railway tracks are captured to identify the distinctive seismic signatures of elephants." DAS applied to elephant detection with precise geolocation.
https://www.bitcommtechnologies.com/distributed-acoustic-sensing

References — Satellite & Remote Sensing

R16WWF / KAZA Study — "4 Million GPS Data Points: WWF Study Reveals Elephant Migration Corridors." Demonstrates GPS-based corridor mapping methodology applicable to Indian Railway risk scoring.
https://www.nathab.com/blog/4-million-gps-data-points-wwf-study-reveals-elephant-migration-corridors
R17EleSense / VFAES — "AI Is Saving India's Elephants from Train Collisions." EleSense thermal imaging + height identification sensor system deploying in Chapramari Wildlife Sanctuary, West Bengal.
https://unchainedtv.com/2024/11/26/how-ai-is-saving-indias-elephants-from-train-collisions/

CLOSING NOTE

GAJA is not a competing system — it is a complementary and additive layer built on the foundation that Indian Railways has already proven works. Every zone where DAS-IDS is operational or sanctioned still lacks predictive intelligence, KAVACH enforcement, and multi-stakeholder data sharing. Every uncovered corridor is a direct deployment opportunity. TRiDE Innovative Technologies is available to present this proposal to the DRM office, zonal Railway headquarters, or MoEFCC Project Elephant division at any time.

CONTACT

Proposal Discussion & Collaboration

Reach out to explore pilot deployment, technical deep-dives, or Railway partnership discussions

Shashi Kanth Poosala

Project Chief — Tride Mobility

Strategic Alliances & Collaborations | Business & Tech Liaison

AI + IoT for Aviation • Railways • Energy • Mobility

📞 +91 94941 56051

✉️ [email protected]

🌐 www.tridemobility.com