From what to build to how to build it — our platform helps you understand, plan, and solve SIH 2026 problem statements with complete insights into requirements, solution approaches, technology stacks, architecture, AI/ML, and implementation roadmaps.
Search for your problem statement, and start building your solution.
| Description: | Background Free Space Optical Communication (FSOC) offers unprecedented advantages for next-generation mobile networks, including gigabit-to-terabit data rates, license-free spectrum operation, high immunity to electromagnetic interference, etc. However, deploying FSOC links between mobile platforms (satellites, UAVs presents a severe challenge of pointing, acquisition and tracking (PAT) of highly narrow laser beams. PAT typically happens in two stages: coarse alignment and fine alignment. Coarse alignment is one of the key challenges of PAT, where the transmitting terminal must first locate and maintain the remote terminal within its camera Field-of-View (FOV). Developing and testing such algorithms on real hardware requires expensive cameras, pan-tilt mechanisms, and optical components & equipment. A software based virtual camera tracking provides an inexpensive and accessible platform for algorithm development and learning. Description Unlike conventional radio-frequency systems, FSOC relies on a highly directional optical beam. Even a small angular error can prevent successful communication. Before fine pointing mechanism can take over, a coarse alignment stage must: • Observe the surrounding environment, • Acquire and detect the remote terminal or beacon, • Estimate the position, and • Continuously adjust the pointing direction to maintain visibility. The participants shall develop this coarse alignment process in software, allowing to develop and validate tracking algorithms without specialized hardware and setup. The following section provides reference parameters and performance criteria to be considered for the software development. Parameters and Specifications Functional Objective: Develop a software system that autonomously detects, identifies, and continuously tracks a designated moving target within a virtual scene by controlling a virtual camera viewport. Add 'Parameters and Specifications' table here Expected Solution Participants shall develop an AI-assisted camera tracking system capable of automatically detecting and continuously tracking a moving optical beacon in a simulated video stream while controlling a virtual pan-tilt camera. The developed software shall be able to: • Generate a configurable virtual environment, • Generate one or more moving targets, • Implement a movable virtual camera, • Detect the target beacon automatically, • Track the beacon continuously using computer vision, • Control and reposition the virtual camera, • Generate and introduce disturbances due to atmospheric turbulence, platform vibrations, camera motion, noise, etc., in the virtual camera feed, • Display tracking performance and statistics in real-time Deliverables Each participating team shall submit the following mandatory deliverables: Software Application A standalone executable application implementing the complete virtual camera tracking system. The application shall provide all the mandatory functions and features as described above. Source Code Complete source code with proper documentation. The code shall be modular and adequately commented. Technical Report The technical report (about 10-15 pages) containing problem understanding, system architecture, description of software modules, tracking methods, AI methods (if used), test methodology, performance analysis and future improvements shall be submitted. User Manual The user manual with the description of installation of software, application operation, parameter configuration, GUI description, etc. shall be submitted. A 3–5 minutes video may also be provided as an optional deliverable for demonstration of the application. Performance Log The software should be capable of automatically generating a performance report containing simulation duration, FPS, acquisition time, average and maximum tracking error, lock retention rate, processing time, etc. Evaluation Method and Criteria The solutions developed by participating teams will be evaluated using multi-layered evaluation method. The following table describes stages of evaluation, their weightage and methods. Add 'Evaluation Method and Criteria' table here |
| Organization: | Indian Space Research Organisation(ISRO) |
| Department: | Department of Space / Indian Space Research Organisation |
| Category: | Software |
| Theme: | Smart Automationtd> |
| Youtube Link: | Not provided |
| Dataset Link: | Additional Information Regarding PS https://drive.google.com/file/d/1AWRWChSMKU8FI38XxfFyQJfCRp3gqkF6/view?usp=drive_link |
| Contact info: | Not provided |
Free Space Optical Communication (FSOC) provides high-bandwidth, license-free, and jam-resistant point-to-point wireless connectivity for satellite-to-satellite, satellite-to-ground, and airborne (UAV/HAPS) terminals [KNOWN]. However, establishing and maintaining an optical link requires ultra-precise Pointing, Acquisition, and Tracking (PAT) due to narrow optical beam divergence [RESEARCHED] (MDPI Aerospace). Developing PAT algorithms directly on physical hardware requires expensive optical benches, fast steering mirrors, pan-tilt gimbals, and high-speed sensors [PS].
This proposal presents OptiTrack-SIM, an end-to-end AI-assisted virtual camera tracking software application for FSOC coarse alignment. It provides a configurable 2D/3D synthetic optical scene, real-time disturbance injection (atmospheric turbulence, platform vibrations, sensor noise), closed-loop virtual pan-tilt camera control, multi-algorithm target detection/tracking (YOLOv8-Nano, SiamRPN, Lucas-Kanade + EKF), and an automated performance evaluation dashboard [INFERENCE].
| Metadata Field | Value |
|---|---|
| PS Number | SIH26169 [PS] |
| Title | Development of an AI-Based Virtual Camera Tracking System for Coarse Alignment of Mobile Free Space Optical Communication (FSOC) Terminals [PS] |
| Organization | Indian Space Research Organisation (ISRO) [PS] |
| Department | Department of Space / Indian Space Research Organisation [PS] |
| Category | Software [PS] |
| Theme | Smart Automation [PS] |
| Deadline | 30 September 2026 [PS] |
| Ideas Counter | 0/500 [PS] |
| Dataset Link | Additional Information Regarding PS (https://drive.google.com/file/d/1AWRWChSMKU8FI38XxfFyQJfCRp3gqkF6/view?usp=drive_link) [PS] |
Fundamental System Objective: To deliver a standalone, physics-informed software simulation suite that models mobile optical beacon trajectories and atmospheric/platform disturbances, executes real-time AI-based target detection and closed-loop virtual camera control for coarse alignment, and outputs automated tracking performance statistics [INFERENCE].
The explicit problem description provided by ISRO requires breaking down the coarse alignment process into software-based simulation modules [PS].
[PS].[PS].[PS].[PS].[PS].[PS].[PS].[PS].[PS][PS][PS][PS][PS][INFERENCE][INFERENCE].[RESEARCHED] (Optica Publishing Group).[INFERENCE].[INFERENCE].| Stakeholder | Goal | Current Pain | System Need |
|---|---|---|---|
ISRO Optical Communication Scientists [INFERENCE] |
Evaluate PAT coarse alignment algorithms under controlled, reproducible disturbance conditions. | Physical testbeds with gimbals, lasers, and optical benches are expensive, slow to reconfigure, and subject to weather [PS]. |
A repeatable, parameter-driven software testbed that models real-world optical channel disturbances and outputs quantitative tracking metrics [INFERENCE]. |
PAT Algorithm Developers [INFERENCE] |
Rapidly benchmark, tune, and test AI detection, state estimation, and control loops. | Hard to isolate algorithm errors from mechanical imperfections in physical hardware [INFERENCE]. |
Modular API/software architecture where detection, tracking, and control logic can be swapped out and benchmarked [INFERENCE]. |
SIH Evaluation Panel / Judges [INFERENCE] |
Verify that the student solution meets all explicit deliverables and mathematically demonstrates coarse alignment. | Many submissions present static video playback or generic bounding box demos without closed-loop camera control [INFERENCE]. |
A live, interactive GUI with real-time parameter tweaking sliders, dynamic camera motion feedback, and downloadable log files [INFERENCE]. |
Field Test Engineers [INFERENCE] |
Transition software-validated tracking algorithms onto actual physical embedded hardware (gimbals/DSP). | Software models often use idealized assumptions that fail under real hardware constraints [INFERENCE]. |
Standardized control output interfaces (e.g., yaw/pitch delta angles) that mimic real servo/gimbal motor commands [INFERENCE]. |
| Feature / Module | Function Description | Justification & Requirement Origin |
|---|---|---|
| Procedural Synthetic Scene Generator | Generates 3D sky/space environments, solar/cloud background noise, and optical beacon geometry. | Directly satisfies Requirement #1 (Configurable Virtual Environment) & #2 (Target Generation) [PS]. |
| Disturbance Injection Pipeline | Applies scintillation intensity fading, beam wander gaussian shifts, high-frequency angular jitter, and Poisson camera shot noise. | Directly satisfies Requirement #7 (Generate and introduce disturbances) [PS]. |
| Virtual Pan-Tilt Viewport Controller | Simulates 2-DOF rotational gimbal kinematics (yaw and pitch limits, angular velocity, inertia). | Directly satisfies Requirement #3 (Movable Virtual Camera) & #6 (Reposition Virtual Camera) [PS]. |
| AI Target Detection & Tracker (YOLOv8-Nano + SiamRPN) | Detects the beacon optical center in the image frame and tracks its bounding box/centroid across frames. | Directly satisfies Requirement #4 (Detect target beacon) & #5 (Track continuously) [PS]. |
| Extended Kalman Filter (EKF) Estimator | Predicts target state during temporary optical blackout or high atmospheric fading. | Ensures robust lock retention, satisfying Requirement #5 & performance logging criteria [INFERENCE]. |
| Real-Time PySide6 / PyQt Dashboard | Displays visual viewport overlay, live tracking error graphs, acquisition time, lock retention rate, and FPS. | Directly satisfies Requirement #8 (Display performance in real-time) & Deliverable #5 (Performance Log) [PS]. |
| Automated Benchmark Report Exporter | Saves CSV/JSON logs containing frame-by-frame error statistics, total lock retention rate, and acquisition time. | Directly satisfies Mandatory Deliverable #5 (Performance Log) [PS]. |
[PS]. The entire system must run on standard consumer workstation hardware (Windows/Linux PC with standard CPU/GPU) [INFERENCE].[INFERENCE].[RESEARCHED] (SPIE Digital Library). The system must adopt optimized, physically verified analytical approximations (Gamma-Gamma intensity fading and Gaussian beam wander) [RESEARCHED].[INFERENCE].[RESEARCHED][INFERENCE][RESEARCHED][RESEARCHED] (MDPI Aerospace).| Data Type | Purpose | Ideal Real-World Source | Hackathon Synthetic Substitute | Access Method | Availability | Privacy Sensitivity | Confidence |
|---|---|---|---|---|---|---|---|
| Target Trajectories | 3D motion paths for moving optical beacons (satellite orbits, UAV waypoint navigation). | Real telemetry logs (SGP4 TLE, UAV flight logs). | Synthetic 3D parametric spline generators (Helical, Lissajous curves, linear acceleration). | On-the-fly Python mathematical generation. | 100% Available | None | [INFERENCE] |
| Atmospheric Turbulence Parameters | Modeling refractive index structure constant $C_n^2$, Fried parameter $r_0$, and scintillation index $\sigma_I^2$. | Atmospheric sensor measurements / LIDAR optical link tests. | Mathematical Gamma-Gamma distribution sampling & von Kármán phase screen overlays. | Pre-calculated lookup tables + online procedural noise generation. | 100% Available | None | [RESEARCHED] |
| Platform Vibration / Jitter Time-Series | Simulating mechanical vibrations from satellite reaction wheels or UAV motor dynamics. | Accelerometer / IMU high-frequency logs from aerospace platforms. | Band-limited White Gaussian Noise (WGN) passed through second-order Butterworth low-pass filters. | Procedural NumPy signal generation script. | 100% Available | None | [RESEARCHED] |
| Synthetic Video Stream | Rendering background sky, sun glare, clouds, and target beacon spot. | High-speed IR / visible optical camera feeds from PAT testbeds. | Real-time OpenCV / PyOpenGL procedural video frame generator. | Direct GPU frame buffer memory array. | 100% Available | None | [INFERENCE] |
| Reference ISRO PS Specs | Problem criteria, metrics, and parameters. | ISRO Google Drive link provided in PS description [PS]. |
Direct parsing of supplied PS text and attached drive link [PS]. |
Direct Google Drive URL parsing [PS]. |
Public PS Link | Public | [PS] |
| Comparable System | Description | Strengths | Weaknesses | Relevance | Remaining Gap | Confidence |
|---|---|---|---|---|---|---|
| MATLAB / Simulink Optical PAT Toolbox | Academic/industrial tool for optical system simulation and control design. | High fidelity wave-optics math and control system toolboxes. | Expensive commercial licensing; high computational overhead; complex UI; not a standalone distribution app. | High theoretical relevance for physics modeling. | Lacks an out-of-the-box lightweight standalone executable GUI with integrated lightweight AI tracking loops [INFERENCE]. |
[RESEARCHED] |
| Gazebo / ROS2 Camera Gimbal Simulators | Open-source robotics simulation framework for drone and camera tracking. | Physics-engine integration, modular ROS2 node structure. | Heavy installation footprint; difficult to deploy as a simple desktop executable; weak optical turbulence modeling. | Relevant for gimbal kinematics and camera viewport control. | Does not model optical channel scintillation, beam wander, or export standardized FSOC PAT benchmark logs [INFERENCE]. |
[RESEARCHED] |
| Standard OpenCV Video Object Trackers | Open-source implementations of KCF, CSRT, GOTURN, and MOSSE tracking algorithms. | High computational efficiency, simple API integration. | Fails under severe scintillation, target occlusion, or intense background solar glare; no integrated gimbal control logic. | High utility for basic feature tracking building blocks. | Unassisted classical trackers drift severely under high atmospheric disturbance without predictive state filtering [RESEARCHED]. |
[RESEARCHED] |
[INFERENCE].[RESEARCHED].[RESEARCHED] (SPIE Digital Library).| Evaluation Criterion | Approach 1: Classical CV + PID | Approach 2: Deep Learning + EKF + PID (Recommended) | Approach 3: Deep RL (PPO/SAC) |
|---|---|---|---|
| Problem Fit | Moderate | High | Moderate |
| Detection Accuracy | Low under noise | Very High | Moderate-High |
| Fading / Occlusion Resilience | Very Low | High (EKF Extrapolation) | Moderate |
| Explainability & Control Stability | High | High | Low |
| Real-time Latency (<15ms) | Very Low (<3ms) | Low (5-10ms) | Low (8-12ms) |
| Hackathon Build Feasibility | 100% | 95% | 50% |
| ISRO Benchmarking Value | Low | Very High | High |
Selected Strategy: Approach 2 — Hybrid AI Visual Perception (YOLOv8-Nano / SiamRPN) + Extended Kalman Filter (EKF) State Estimation + Predictive Pan-Tilt Controller.
[RESEARCHED]. Deep neural networks (YOLOv8-Nano) retain feature representations under severe blurring/scintillation, while the EKF provides motion prediction during momentary blackout frames [INFERENCE].[INFERENCE]. Combining AI perception with a deterministic EKF + PID controller guarantees stable kinematic convergence and predictable camera repositioning.[KNOWN], ensuring smooth real-time execution alongside the simulation physics renderer.[INFERENCE].[INFERENCE].+-----------------------------------------------------------------------------------+
| PRESENTATION / GUI LAYER |
| PySide6 Desktop Application | Real-Time Viewport Overlay | Parameter Sliders |
| Live KPI Graphs (Error, FPS, Lock %) | Performance Log Export Button |
+-----------------------------------------------------------------------------------+
|
User Actions / Configurations
v
+-----------------------------------------------------------------------------------+
| SIMULATION & PHYSICS ENGINE |
| 3D Kinematic Target Engine | Disturbance Generator (Scintillation, Jitter) |
| Virtual Camera Viewport Kinematics | Frame Renderer (OpenCV / PyOpenGL) |
+-----------------------------------------------------------------------------------+
|
Synthetic Video Frames (Image Array)
v
+-----------------------------------------------------------------------------------+
| AI PERCEPTION & TRACKING ENGINE |
| YOLOv8-Nano / SiamRPN Object Detector --> Centroid Measurement (u_v) |
| Extended Kalman Filter (EKF) State Estimator --> State Vector [x, y, vx, vy, ax, ay]|
+-----------------------------------------------------------------------------------+
|
Target Angle Centroid Error (Azimuth, Elevation)
v
+-----------------------------------------------------------------------------------+
| CLOSED-LOOP CAMERA CONTROLLER |
| Feed-Forward PID Pan-Tilt Controller | Angular Velocity Limits & Motor Deadband |
| Output: New Camera Orientation Angles (Theta_Az, Theta_El) |
+-----------------------------------------------------------------------------------+
|
Feedback Control Loop
v
+-----------------------------------------------------------------------------------+
| ANALYTICS & EVALUATION ENGINE |
| Acquisition Time Clock | Instantaneous & RMS Tracking Error Calculator |
| Lock Retention Rate (%) | FPS Monitor | Standalone Report Exporter (JSON/CSV) |
+-----------------------------------------------------------------------------------+
| Layer | Selected Technology | Why It Fits | Alternative Considered | Why Alternative Was Rejected |
|---|---|---|---|---|
| GUI Framework | PySide6 (Qt for Python) | Cross-platform, hardware-accelerated QOpenGLWidget for video rendering, rich UI controls, smooth 60 FPS performance [KNOWN]. |
Tkinter / Streamlit | Streamlit cannot handle real-time 60 FPS video loop rendering; Tkinter is visually outdated and lacks smooth graphing controls [INFERENCE]. |
| Simulation Backend | Python 3.12 + NumPy + OpenCV | High computational speed for matrix math, seamless integration with OpenCV and PyTorch PyInstaller builds [KNOWN]. |
C++ / DirectX | C++ increases GUI development overhead significantly for a hackathon timeline [INFERENCE]. |
| AI Perception | YOLOv8-Nano (ONNX Runtime) | Ultra-lightweight (~6MB weights), lightning-fast inference (<5ms on CPU via ONNX execution provider) [RESEARCHED]. |
Heavy ResNet / YOLOv8-X | High latency (>40ms on CPU) breaks real-time closed-loop control constraints [INFERENCE]. |
| State Estimation | FilterPy (Extended Kalman Filter) | Flexible Python library for EKF state estimation, motion models, and covariance tuning [KNOWN]. |
Custom C++ Filter | Unnecessary complexity for 2D/3D kinematic state prediction [INFERENCE]. |
| Data Logging & Export | Pandas + Matplotlib / PyQtDataVisualization | Clean tabular log generation, simple export to CSV/JSON, smooth real-time plotting [KNOWN]. |
SQLite Database | Overhead of database setup for simple session-based performance log exporting [INFERENCE]. |
| Packaging / Distribution | PyInstaller / Nuitka | Compiles Python code and Qt dependencies into a single standalone .exe or executable directory [KNOWN]. |
Docker Container | Docker requires host virtualization and setup overhead for ISRO judges during evaluation [INFERENCE]. |
[PS].[RESEARCHED].[INFERENCE].[RESEARCHED].[INFERENCE].[INFERENCE].[INFERENCE].[INFERENCE].[INFERENCE].[PS].[INFERENCE].[RESEARCHED].[INFERENCE].[INFERENCE].[INFERENCE].config.json){
"environment": {
"resolution": [1280, 720],
"fps": 60,
"background_type": "deep_space",
"solar_glare_intensity": 0.2
},
"target": {
"trajectory_type": "lissajous",
"base_velocity_m_s": 15.0,
"beacon_power_mw": 100
},
"disturbances": {
"cn2_turbulence": 1e-14,
"vibration_freq_hz": 12.0,
"vibration_amplitude_pix": 8.5,
"shot_noise_sigma": 0.05
},
"camera_gimbal": {
"fov_deg": 15.0,
"max_pan_speed_deg_s": 30.0,
"max_tilt_speed_deg_s": 30.0,
"pid_gains": {"kp": 0.8, "ki": 0.05, "kd": 0.12}
}
}
performance_log.json){
"session_metadata": {
"timestamp": "2026-09-30T10:15:00Z",
"simulation_duration_sec": 60.0,
"total_frames_rendered": 3600
},
"summary_kpis": {
"average_fps": 59.4,
"acquisition_time_sec": 0.82,
"mean_tracking_error_deg": 0.042,
"max_tracking_error_deg": 0.185,
"lock_retention_rate_pct": 98.6,
"avg_processing_latency_ms": 7.2
}
}
[INFERENCE].[INFERENCE].[INFERENCE].[INFERENCE].[INFERENCE].[PS][PS].[PS].[PS].[PS].[PS].[PS].[PS]..exe binary, complete documented source code, Technical Report (10–15 pages), User Manual [PS].[INFERENCE].[INFERENCE].[PS].[PS].[INFERENCE].[RESEARCHED].To present a compelling demo to ISRO judges without physical optical hardware, the application includes a Live Interactive Simulation Dashboard:
+------------------------------------------------------------------------------------+
| OPTITRACK-SIM: ISRO FSOC VIRTUAL CAMERA TRACKING SYSTEM |
+--------------------------------------------------+---------------------------------+
| VIRTUAL CAMERA VIEWPORT [60 FPS] | LIVE DISTURBANCE CONTROLS |
| +----------------------------------------------+ | Turbulence (Cn2): [---||---] |
| | Target | | Vibration Amp: [-----||-] |
| | (o) [Lock] | | Target Speed: [---||---] |
| | + | | Camera FOV: [--||----] |
| | Viewport Center | +---------------------------------+
| | | | REAL-TIME METRIC DASHBOARD |
| +----------------------------------------------+ | FPS: 59.8 | Latency: 6.4 ms |
| [Status: COARSE ALIGNMENT LOCKED] | | Acq Time: 0.74 s |
+--------------------------------------------------+ | Mean Error: 0.038 deg |
| REAL-TIME TRACKING ERROR PLOT (DEG VS TIME) | | Lock Retention: 99.2% |
| 0.2 |~~~~~~~~~~~~~~~~ | +---------------------------------+
| 0.1 | \ | | ACTIONS |
| 0.0 +-----------------\---------------------- | | [START] [INJECT FADE] [EXPORT] |
+--------------------------------------------------+---------------------------------+
[PS])2026 Phase Timeline:
[May - Jun 2026] Phase 1: Mathematical Foundations & Disturbance Engine
[Jul 2026] Phase 2: AI Perception Training & EKF State Estimator
[Aug 2026] Phase 3: Closed-Loop Gimbal Kinematics & PySide6 GUI
[Sep 1 - 15] Phase 4: System Integration, Benchmarking & Log Exporter
[Sep 16 - 30] Phase 5: Technical Report (15 pages), User Manual & Executable Build
[INFERENCE].[INFERENCE].[INFERENCE].[INFERENCE].[PS].| Test Category | Target Benchmark / Metric | Test Procedure | Pass / Fail Criteria |
|---|---|---|---|
| Functional Test | Target Detection & Locking | Spawn target at random FOV coordinates. | System detects target within <1.0 second and centers viewport [INFERENCE]. |
| Performance Test | Frame Rate & Processing Latency | Run simulation at $1280 \times 720$ resolution on standard i5 CPU / GTX 1650 GPU. | Frame rate $\ge 50\text{ FPS}$; processing latency per frame $\le 15\text{ ms}$ [INFERENCE]. |
| AI Detection Accuracy | Object Detection mAP | Evaluate YOLOv8-Nano on 500 test synthetic frames with clouds/glare. | $\text{mAP@50} \ge 0.92$; Precision $\ge 0.90$ [INFERENCE]. |
| Disturbance Robustness | Scintillation & Fading Test | Apply extreme turbulence ($C_n^2 = 1 \times 10^{-13}$) causing 70% intensity drop. | Lock Retention Rate $\ge 90%$ over a 60-second run [INFERENCE]. |
| False Positive Resilience | Background Glare Distraction | Introduce solar reflection artifacts near the target beacon. | Detector maintains tracking on true optical beacon without jumping [INFERENCE]. |
| Log Exporter Integrity | Performance Log Generation | Execute 60s run and export log file. | JSON file correctly records FPS, acquisition time, average error, and lock retention [PS]. |
| Risk Event | Likelihood | Impact | Severity | Mitigation Strategy |
|---|---|---|---|---|
| AI Detection Overhead Slows Rendering Below 30 FPS | Moderate | High | High | Export YOLOv8-Nano model to ONNX Runtime format with CPU thread optimization; downsample AI input matrix to $320 \times 320$ [INFERENCE]. |
| PID Camera Controller Oscillates Under High Vibration | Moderate | Moderate | Medium | Implement derivative filtering and Extended Kalman Filter state smoothing before feeding errors to the PID loop [INFERENCE]. |
| PyInstaller Executable Fails on Judges' Laptops Due to Missing DLLs | Low | High | High | Perform automated multi-OS CI/CD build testing on clean Windows 10/11 sandbox environments [INFERENCE]. |
| Unrealistic Disturbance Math Discredited by ISRO Experts | Low | High | High | Ground all scintillation and jitter equations directly in peer-reviewed FSOC literature (Gamma-Gamma and von Kármán power spectrums) [RESEARCHED]. |
[INFERENCE].[INFERENCE].[INFERENCE].If ISRO transitions this software simulation suite into an industrial PAT testbed:
[KNOWN].[RESEARCHED].[RESEARCHED].| Parameter | Generic Student Submission | OptiTrack-SIM Proposed Solution |
|---|---|---|
| Simulation Realism | Static video file playback or basic bouncy ball graphics. | Physics-informed optical channel simulation (Gamma-Gamma scintillation fading, Gaussian beam wander, low-pass platform jitter) [RESEARCHED]. |
| Tracking Mechanism | Simple OpenCV color bounding box without closed-loop feedback. | Hybrid YOLOv8-Nano + Extended Kalman Filter (EKF) providing state prediction during total optical blackout [INFERENCE]. |
| Control System | Hardcoded camera movement commands. | 2-DOF Kinematic Gimbal Controller modeling motor speed limits, angular acceleration, and closed-loop PID control [INFERENCE]. |
| Evaluation Metrics | Manual visual inspection. | Automated ISRO Benchmark Log Exporter outputting exact metrics specified in PS (Acquisition Time, Lock Retention %, Mean Error, FPS) [PS]. |
[INFERENCE].[RESEARCHED].[RESEARCHED] (MDPI Aerospace).[ASSUMPTION][ASSUMPTION-1]: Software-only evaluation implies no physical motor hardware testing is expected during SIH judging.[ASSUMPTION-2]: Standard consumer laptops (Intel i5/i7, 8GB RAM, integrated/dedicated GPU) represent the target execution environment for the standalone .exe.[ASSUMPTION-3]: The optical beacon can be approximated in 2D image frames as a Gaussian intensity spot with dynamic brightness and spatial centroid offsets.[NEEDS VALIDATION][NEEDS VALIDATION-1]: Exact numerical ranges for ISRO's preferred camera FOV (e.g., 5° vs 15°) and gimbal angular speed limits (to be verified via attached Drive link or mentor interaction).[NEEDS VALIDATION-2]: Weightage assigned to optional 3-5 minute demonstration video during preliminary screening.Build OptiTrack-SIM as a standalone PySide6 desktop executable featuring an interactive 3D synthetic optical environment, physics-grounded disturbance generators (turbulence/jitter), a hybrid YOLOv8-Nano + EKF computer vision tracking loop, and an automated ISRO KPI benchmark log exporter [INFERENCE].
.exe)Synthetic image dataset quality and realistic disturbance mathematical models (scintillation and beam wander) [INFERENCE].
Inference latency spikes on non-GPU laptops dropping simulation frame rates below 30 FPS during judging [INFERENCE].
Simulating computationally realistic atmospheric turbulence fading and beam wander in real-time while maintaining 60 FPS video loop rendering [RESEARCHED].
[PS].[INFERENCE].[RESEARCHED].[INFERENCE].[INFERENCE].Verify ONNX Runtime inference latency on CPU (<8ms) for $320 \times 320$ frame sizes to ensure 60 FPS execution without requiring a dedicated NVIDIA GPU [INFERENCE].
92 / 100 (Software-only scope removes expensive hardware dependencies; standard computer vision and control libraries make full execution highly achievable within the project timeline) [INFERENCE].
HIGH [INFERENCE]
MDPI Aerospace / Photonics Review of Pointing Modules and Gimbal Systems for Free-Space Optical Communication
MDPI Electronics - Machine Learning-Based Beam Pointing Error Reduction for Satellite-Ground FSO Links
Optica Publishing Group - Beam Wander Prediction with Recurrent Neural Networks & Turbulence Modeling
SPIE Digital Library - Deep Reinforcement Learning for Beam Pointing Control in Laser Communications
SIH 2026 Official Portal Problem Statement Link (ISRO Additional Dataset Link)
[PS].What we think about this platform, and how it can help people solve SIH problems.
" 30M+ Google Gemini tokens went into researching, analyzing, and understanding all 233 SIH problem statements to make them easier to explore and solve. "
" The platform makes SIH problem statements much easier to understand. Having everything from the problem details to the procedure in one place. "
" Oh, you got ads? Well, we need to make money too! But, you can use our platform to help you win SIH and take home the prize money. "
Get in touch if you need help bringing your ideas to life, or simply have an interesting idea you'd like to talk through. I'm always open to a conversation and exploring new possibilities together.