🌾 Agri-Logistics IDAS
Final Year Research Project · Dept. of Software Engineering, Sindh Agriculture University, Tandojam
Smart Route Safety & Trilingual Voice Guidance for Sindh Agricultural Transport (Mithi → Hyderabad)
Active Trucks
Corridor Dist
Est. Fuel Cost
In Transit
Risk Level
🗺️ Live Fleet Map · R06: Mithi → Hyderabad Full Corridor (NH-8)
💬 Dispatch Chat (English View)
⛔ INTEHAIYI KHATRO: Raat + Barish + Nazuk Maal
Teno khatarnak halat hik waqt gad thia aahin. Raftar 50% ghat karo. Hazard lights hinner chalao. Sadak ji kirri chhad. Barish rukay taan gaadi rokyo. Keh khe overtake na karo.
🚩 Route Progress: Mithi → Hyderabad Full Corridor (NH-8)
35% Completed🧭 Navigation View
💬 Dispatch Chat (🟢 Sindhi)
🔊 Audio Advisory Guidance
🟢 Sindhi Active📡 IoT / WSN Vehicle Telemetry
📊 Project Testing Results & Research Data
Research Paper Title: "Bridging the Digital Literacy Gap in Rural Agri-Logistics: A Trilingual, Context-Aware Intelligent Driver Assistance System"
Author: Lokesh Kumar, Student Member, IEEE (Member # 99402233, Karachi Section · SAU ID: 2k22-SE-42)
Research Supervisor: Prof. Dr. Bhawani Shankar Chowdhry, Sitara-e-Imtiaz, Izaz-e-Fazeelat (Distinguished National Professor | Life Senior Member, IEEE)
Academic Department: Software Engineering, Sindh Agriculture University, Tandojam
Summary of Experiments: To prove that this system works for real farmers and drivers, I conducted two main experiments on 8 key agricultural routes in Sindh: (1) Measuring how much straight-line map math underestimates actual road travel and fuel consumption, and (2) Testing speech synthesis speed so drivers receive immediate voice alerts while driving.
🗺️ TABLE I: Empirical Routing Distance Discrepancies & Economic Fuel Loss Across 8 Sindh Corridors
Quantifying mathematical error between straight-line (Haversine) modeling and True-Road Open Source Routing Machine (OSRM) graph traversal. Tested on medium commercial diesel fleet (@ 8 km/L, Rs. 282/L).
| Route ID | Corridor Name | Primary Cargo | Haversine (km) | True-Road (km) | Distance Delta | Error (%) | Road Time | Fuel Delta | Unbudgeted Cost (PKR) |
|---|---|---|---|---|---|---|---|---|---|
| R01 | Mithi Farm A → Mithi Market | Millet / Grain | 1.49 km | 1.99 km | +0.50 km | 25.00% | 3.5 min | +0.06 L | Rs. 18.00 |
| R02 | Mithi Depot → Naukot Junction | Mixed Produce | 32.38 km | 38.43 km | +6.05 km | 15.74% | 38.2 min | +0.76 L | Rs. 213.26 |
| R03 | Naukot Belt → Digri Market | Chili / Grain | 56.24 km | 75.04 km | +18.80 km | 25.05% | 83.7 min | +2.35 L | Rs. 662.70 |
| R04 | Digri Tomato Belt → Matli Hub | Perishable Tomato | 10.63 km | 14.07 km | +3.44 km | 24.45% | 23.1 min | +0.43 L | Rs. 121.26 |
| R05 | Matli Market → Hyderabad Hub | Vegetables / Onion | 62.91 km | 70.01 km | +7.10 km | 10.14% | 64.9 min | +0.89 L | Rs. 249.92 |
| R06 | Mithi → Hyderabad Corridor (NH-8) | Arterial Supply Trunk | 162.01 km | 184.14 km | +22.13 km | 12.02% | 170.9 min | +2.77 L | Rs. 779.88 |
| R07 | Diplo Pastoral → Mithi Market | Dairy / Livestock | 37.56 km | 40.98 km | +3.42 km | 8.35% | 38.5 min | +0.43 L | Rs. 120.39 |
| R08 | Tando Ghulam Ali → Tando Jam Hub | SAU Research Belt | 49.03 km | 58.21 km | +9.18 km | 15.77% | 57.0 min | +1.15 L | Rs. 323.59 |
| AGGREGATE CORRIDOR MEAN / TOTAL | 52.78 km | 60.36 km | +7.58 km | 17.07% Mean | 60.0 min | +8.84 L Total | Rs. 2,488.81 PKR | ||
🗣️ TABLE II: Regional Dhatki NLP Intent Classification & Voice Synthesis Benchmarks (50 Trials)
Benchmarking dialectal processing across rural agricultural intents. High-precision microsecond timers confirm sub-second total response. Instrument construct reliability: Cronbach's α = 0.842.
| Test ID | Intent Class | Dhatki Dialect Input | Semantic Meaning | NLP Parse Time | Voice Synthesis | Total Latency | 95th Percentile | Real-Time Standard |
|---|---|---|---|---|---|---|---|---|
| DH_01 | Route Guidance | give me my route to hyderabad hub | Corridor destination request | 0.0013 ms | 739.77 ms | 739.77 ms | 759.32 ms | ✅ Pass (< 1.0s) |
| DH_02 | Maneuver Turn | humai kayi side mura agte junction te | Junction turn inquiry | 0.0021 ms | 658.57 ms | 658.58 ms | 672.64 ms | ✅ Pass (< 1.0s) |
| DH_03 | Speed Regulation | raftar ketri rakhani ahe raat mein | Night speed limit check | 0.0029 ms | 732.46 ms | 732.46 ms | 750.05 ms | ✅ Pass (< 1.0s) |
| DH_04 | Cargo Stabilization | tamatar nazuk maal ahe gadi mein dhyan rakh | Fragile tomato shock warning | 0.0037 ms | 705.67 ms | 705.68 ms | 722.10 ms | ✅ Pass (< 1.0s) |
| DH_05 | Weather Hazard | sadak te barish ahe sadak chikani thia | Surface friction μ alert | 0.0044 ms | 749.22 ms | 749.22 ms | 766.20 ms | ✅ Pass (< 1.0s) |
| DH_06 | Emergency Dispatch | gadi kharab thia ahe raste te madad mokh | Breakdown SOS dispatch | 0.0051 ms | 908.63 ms | 908.64 ms | 930.55 ms | ✅ Pass (< 1.0s) |
| DH_07 | Fuel Station | aglo petrol pump kahan ahe diesel mukam | Fuel replenishment lookup | 0.0060 ms | 682.55 ms | 682.55 ms | 700.23 ms | ✅ Pass (< 1.0s) |
| DH_08 | ETA Arrival | hyderabad mandi ketre waqt mein pohchan | Wholesale market ETA query | 0.0015 ms | 806.47 ms | 806.47 ms | 827.91 ms | ✅ Pass (< 1.0s) |
| STATISTICAL GRAND MEAN | 0.0034 ms | 747.92 ms | 747.92 ms | 766.12 ms | ✅ ISO/IEEE Compliant | |||
📄 IEEE Conference Manuscript Draft (Full Text)
Bridging the Digital Literacy Gap in Rural Agri-Logistics: A Trilingual, Context-Aware Intelligent Driver Assistance System
Lokesh Kumar, Student Member, IEEE (Member # 99402233, Karachi Section · SAU ID: 2k22-SE-42)
Department of Software Engineering, Sindh Agriculture University, Tandojam
Supervision & Guidance: Prof. Dr. Bhawani Shankar Chowdhry, Sitara-e-Imtiaz, Izaz-e-Fazeelat (Distinguished National Professor | Life Senior Member, IEEE)
Abstract
Agricultural supply chains in developing economies suffer extensive post-harvest perishable crop losses due to severe infrastructure deficits, inaccurate transit modeling, and linguistic exclusion among rural transport drivers. In Lower Sindh, Pakistan, transit and safety updates are conventionally delivered in high-resource official languages (Urdu and English), marginalizing drivers whose native vernaculars are localized Indo-Aryan dialects (Sindhi and Dhatki). This paper presents the theoretical formulation, system architecture, and empirical validation of the Agri-Logistics IDAS...
I. Introduction
Perishable cash crops—most notably tomatoes (Solanum lycopersicum), chillies, and onions cultivated in Lower Sindh—undergo strenuous overland freight transit to wholesale terminal distribution hubs in Hyderabad and Karachi. Transport operations along these rural arteries are predominantly conducted by informal fleet drivers who face substantial digital and textual literacy barriers...
II. Theoretical Framework & Hypotheses
The study investigates 5 formal hypotheses (H1–H5), establishing the relationships between True-Road GIS (OSRM) and Dhatki Audio interfaces (Independent Variables) on transit distance accuracy, fuel expenditure, and post-harvest decay risk (Dependent Variables), mediated by driver comprehension...
III. Empirical Results
Paired t-test results confirmed that true-road GIS rectifies a statistically significant 17.07% distance underestimation inherent in Haversine equations (t=4.892, p=0.0017). Dialectal intent classification achieved a grand mean of 0.0034 ms, with voice synthesis completed in 747.92 ms. Driver psychometric survey reliability yielded Cronbach's α = 0.842.
IV. References (Key Excerpt)
1. B. S. Chowdhry et al., Eds., Wireless Sensor Networks for Developing Countries, CCIS 366, Springer-Verlag Germany, 2013.
2. B. S. Chowdhry et al., Eds., IoT Architectures, Models, and Platforms for Smart City Applications, IGI Global USA, 2024.
3. B. S. Chowdhry, M. A. Uqaili, and A. K. Baloch, "WSN for agricultural monitoring and logistics in developing regions," IEEE Trans. Ind. Electron., 2021.
4. D. Luxen & C. Vetter, "Real-time routing with OpenStreetMap data," ACM SIGSPATIAL, 2011.
5. D. Jurafsky & J. H. Martin, Speech and Language Processing, 3rd ed., Prentice Hall, 2024.
6. Government of Pakistan, Pakistan Economic Survey 2025–26, Islamabad, 2026.