INDIA | USA | TAIWAN

Indranet

By DefenSys

Mission

Methods in this document form the basis of design of Indranet - a defence system consisting of a multi-static radar network called Radionet, consisting of at least three receivers, and a mesh construction called Buzznet, comprising of possibly thousands of acoustic sensors capable of locating acoustic targets such as drones. Both designs rely on locating the source using the difference in arrival times of wavefronts at various sensors using the Narad algorithm. Utilising a network of sensors instead of a monolithic setup provides several advantages, such as:

• Increased signal-to-noise ratio for wavefronts populating sub-noise levels leading to increased range of detection and earliness in warnings.
• Reduced likelihood of jamming due to increased coverage and spread in terms of surface area.
• Significantly reduced number of points of critical failures due to distributed sensing nodes.
• Significant cost-effectiveness in terms of increased tolerance to noise in individual receivers or sensors at fixed cost.

Due to Narad's ability to detect sub-noise targets, Indranet can also reveal bodies below the horizon by piggybacking on atmospheric channelling. Indranet can also be adapted for Naval use by deploying sensors on sub-surface buoys.

Figure 1: Concept

Introduction

The global race for advanced ordnance, missiles, and drone technologies has put significant pressure on the readiness and efficiency of defenders' systems. Radars are the current best method for detecting missiles, but localising them from a single observation with a monolithic setup is fundamentally impossible. Defenders typically use secondary means, such as multiple observations to measure doppler modulation, to determine precise locations. This document presents an alternative: a multi-static, de-localised network of receivers, which offers significantly higher sensitivity and longer range than monolithic setups at fixed costs.

This draft elaborates on designing a similar network to detect drones by targeting their native acoustic emissions. Methods outlined significantly increase the sensitivity and range of the acoustic sensor network and can be adopted by existing networks. The document further outlines the Narad algorithm for resolving missiles or drones in their respective emission channels and presents three implementations: Radionet, Radionet+, and Buzznet.

Principle

The physical processes underlying Radionet and Buzznet are similar. While Buzznet relies on the native emission from the acoustic source, Radionet utilises the reflected radio pulses from a known friendly transmitter. Both act as pseudo-sources of incoming signals in their respective channels and propagation media. Narad is agnostic to these physical effects and focuses on the digital encoding of the incoming signal.

Signal Lifecycle

1. Emission: The source emits the signal isotropically, traveling to different sensors at unequal times.
2. Reception: Sensors pick up the signals and transmit them in raw digital form to the Narad algorithm.
3. Modulation and Noise: The signal is modulated in amplitude and frequency due to sensor orientation and source motion, and populated with various noises.
4. Digital Encoding: Sensors transmit the noisy, modulated signal to Narad.

Narad's Role

Narad accounts for known noise sources, averages over unknown noise contributions, recovers the signal above threshold SNR, triangulates the source's position from arrival times, demodulates frequency to determine speed and angle, demodulates amplitude, and calculates parameters like mass and count in a swarm.

Narad Algorithm

Narad × Radionet

Figure 2: Radionet

When interfaced with Radionet, Narad initiates with the coherent pulse burst (cPB) algorithm, targeting unmodelled waveforms appearing as short bursts of increased power. The cPB algorithm detects these spikes by coherently combining the SNR from different sensors without needing to know the precise shape or extent of the burst. This involves estimating background noise and processing real data events that exceed the threshold. The coherent noise decreases as the number of sensors increases, and noise cancelling techniques are employed.

Narad × Radionet+

Radionet+ integrates SNR of sub-noise pulses in time domain across sensors to match monolithic receiver sensitivity. This method is effective for signals near the background noise level and can be extended to continuous radio wave transmitters. The analysis for continuous waves uses Wiener filters of precomputed waveforms, cross-correlating raw data with template waveforms and normalising with noise power.

Narad × Buzznet

Figure 3: Buzznet

When interfaced with Buzznet, Narad uses long time-baseline integrators for longer waveforms, such as drone hums. This signal is a long-lasting interference of nearly monochromatic signals, presenting with stationary lines in frequency space. Source localisation is achieved by analyzing phase differences of incoming wavefronts at sensors. Noise decreases with longer integration times, while signal contribution to SNR increases proportionally.

Sensors

The specifications of radio and acoustic sensors are critical in the Indranet infrastructure. Narad's effectiveness is determined by the sensors' parameters, such as false alarm rate for Radionet and noise sensitivity for Buzznet; Lidas is an example of an appropriate acoustic sensor for Buzznet built on the Liffs architecture.

Algorithm Interfacing

Radionet

Radionet relies on a minimum of three radar receivers spaced sufficiently apart to triangulate a source. Realistically, four receivers are preferred to avoid redundancies and account for errors. Differences in arrival times result in hyperbolic surfaces, solving for the source's position. Doppler-modulated pulse repetition frequency measurements allow determination of the source's velocity and acceleration, with additional sensors enabling discarding 'ghost' sources from reflections.

Radionet+

Figure 4: Radionet+ tuned to sub-horizon sources

Radionet+ extends Radionet's capabilities by integrating SNR of sub-noise pulses over time, suitable for detecting distant sources. Continuous wave analysis uses Wiener filters, cross-correlating raw data with template waveforms and normalising with noise power, effectively increasing detection sensitivity. Due to the natural increase in noise-cancellation in a network and the additional spatial degrees of freedom provided by numerous sensors, Radionet+ can also be tuned to triangulate sources far below the horizon.

Buzznet

Buzznet uses long time-baseline integrators tuned to longer waveforms, such as drone hums. Source localisation is achieved by phase difference analysis of incoming wavefronts at sensors. The integration of signals over long timescales decreases noise contribution and increases signal contribution to SNR.

Remarks

The multi-static radar network and acoustic sensor mesh presented here, powered by the Narad algorithm, offer significant advantages over traditional monolithic setups. These include increased sensitivity, range, and robustness against noise and jamming, providing a cost-effective and highly efficient solution for modern defense needs.