0%
INITIALIZING_SYSTEM...
SYSTEM STATUS: ONLINE // V 2.0

NavX // ARN

> AI-DRIVEN INFRASTRUCTURE FOR RESILIENT CITIES.
> EXECUTED VIA ROS 2 + YOLOv8 ON EDGE.

An autonomous mobility kernel designed to bridge the urban data gap. NavX repurposes advanced safety logic to navigate, map, and serve informal settlements and unmapped disaster zones.

Initialize System_
NavX System Visual

// SYSTEM ALERT: FRAGILE INFRASTRUCTURE

Standard navigation stacks optimize strictly for T_min (Velocity). In dense, informal, or disaster-struck environments, this leads to failure.

Current robots are blind to Inequality, Infrastructure Decay, and Human Safety. 1 Billion people live in "unmapped" zones where the "fastest path" is often the most dangerous.

// PATCH APPLIED: RESILIENCE KERNEL

NavX introduces Risk-Aware Path Scoring.

By prioritizing Safety > Velocity, NavX autonomously navigates complex urban terrain, detecting hazards (debris, blocked paths) in real-time.

RESULT: A deployable Digital Twin generator for inclusive city planning.

Core Capabilities

01

Explainable AI

Route transparency via explicit cost functions. Building trust between autonomous agents and vulnerable communities.

02

Dynamic Hazard Avoidance

Automatic re-planning < 200ms upon detecting debris, crowds, or structural risks. Designed for unstructured roads.

03

ROS 2 Native Node

Seamless integration with City Digital Twins. Functions as a decentralized edge node for municipal data collection.

04

Inclusive Metrics

Real-time readout: "−18% Velocity, +40% Safety". Quantifiable trade-offs for emergency logistics.

ALGORITHM_COST_FUNCTION

C = (w_t × Time) + (w_r × Resilience) + (w_i × Inclusivity)

Profile Configuration

MODE T_Weight R_Weight I_Weight DESCRIPTION
RESPONSE 1.0 0.2 0.1 *Formerly Velocity.* Emergency Mode. High speed for rapid intervention/First Responders.
RESILIENCE 0.5 1.0 0.3 *Formerly Security.* Disaster Mode. Max obstacle clearance. Navigates debris & unmapped terrain.
INCLUSIVE 0.4 0.3 0.8 *Formerly Comfort.* Accessibility Mode. Smooth motion for transporting fragile aid or guiding elderly.

PERCEPTION_SYSTEM

YOLOv8 VISION CORE // REAL-TIME INFERENCE

HAZARD ANALYSIS FEED [ACTIVE]
Traffic Analysis Feed
PEDESTRIAN 98%
DEBRIS 95%
SEGMENTATION // CAM_02 [ACTIVE]
Segmentation Feed

> DETECTION_PIPELINE

  • MODEL: YOLOv8 (Custom TensorRT Build)
  • INPUT: 1280x720 @ 60FPS
  • LATENCY: ~18ms (End-to-End)
  • CLASSES: Person, Debris, Flood_Zone, Emergency_Vehicle

> DEPTH_ESTIMATION

Monocular depth approximation using bounding box regression constraints. Calibrated for urban canyon environments.

CALIBRATION_STATUS: NOMINAL
ERROR_MARGIN: < 5% @ 10m

HAZARD_SCORING_FORMULA

Risk = Σ (Confidence × ClassWeight × DistanceFactor)

System Architecture

Distributed nodes. Modular design. Fail-safe.

HARDWARE_LAYER STATUS: ACTIVE RAZER CAM > YOLOv8 NODE HAZARD_SCORING RISK_MATRIX_CALC ● LIVE ROUTER (A*) COST = T(w) + R(w) + C(w) WAYPOINT_CLIENT ROS2 NAV_STACK FASTAPI_GATEWAY WEBSOCKET / REST MAP_VIEW CTRL USER_INTERFACE
> Router_Mod

Python A* implementation. Matrix evaluation.

> Signals_Node

Sensor fusion aggregation point.

> Waypoint_Client

Nav2 interface and action management.

SYSTEM_SPECIFICATIONS

HARDWARE & SOFTWARE INFRASTRUCTURE

> HARDWARE

Razer Kiyo Pro
TurtleBot3 Burger
Raspberry Pi 4 (4GB)

> AI_VISION

YOLOv8n (Nano)
60FPS Inference
Hazard Scoring Engine

> CORE_ROBOT

ROS 2 Humble
Nav2 Stack
Python 3.10

> DASHBOARD

React.js v18
TailwindCSS
FastAPI / WebSocket

SYSTEM_PERFORMANCE

REAL-TIME METRICS & LATENCY ANALYSIS

INFERENCE TIME
18-25ms
FPS
30-60FPS
REPLANNING
SPEED
200-500ms
DETECTION ACCURACY
~90%
INTERACTIVE_DEMO // V.1.4

Path Planning Visualization

Select a routing profile to observe real-time solver behavior.

HAZARD START TARGET
STATUS: MULTI_PATH_SIMULATION
■ VELOCITY
■ SECURITY
■ COMFORT
// VELOCITY_PROFILE
Optimizes for T_min. Accepts higher risk thresholds for direct routing.
RISK:
HIGH
[] SECURITY_PROFILE
Maximizes obstacle clearance. Reroutes around all detected hazard zones.
RISK:
LOW
~~ COMFORT_PROFILE
Minimizes lateral acceleration (G-force). Smooth curves only.
RISK:
MED
PERSONNEL_DATABASE // ACESS_LEVEL_1

Command Unit

ACTIVE_AGENTS: 09
SYNC_STATUS: 100%