CPSI — Coded Propagation-Steered Illumination · US 2026/0056318 A1 · Priority October 2024

One guided medium.
Two orthogonal ways to encode direction.

The SciPhAI CPSI architecture supports two independent operating modes on identical hardware: TDE — Temporal Direction Encoding — the primary mode, where propagation delay encodes direction, requiring only direct detection — and CFSRP — Continuous Frequency-Shifting Radiating Pulse — the secondary mode, adding velocity measurement and longer range via coherent detection.

HIGH FREQUENCY (BLUE) → → → → → LOW FREQUENCY (RED) — one pulse, one fiber, all directions

Live Physics Simulation

Watch the pulse propagate —
frequency shifting in real time

The animation below simulates a single optical pulse traveling through a nonlinear waveguide coil. The pulse enters at the left end with a high carrier frequency (blue). As it propagates, nonlinear effects (soliton self-frequency shift via stimulated Raman scattering) cause the carrier frequency to continuously decrease — shifting through cyan, green, amber, and into red. At each aperture along the coil, light radiates outward into a specific direction, carrying the frequency the pulse has reached at that exact position.

CFSRP Waveguide Propagation Simulator
SPEED
MODE: CFSRP — Direction encoded in carrier frequency via Raman shift ν(z) = ν₀ − α·z
Pulse Position
0.00m
Carrier Frequency
193.4THz
Active Aperture
Illumination Angle
°
Real-Time Frequency Spectrum — Current Pulse State (CFSRP Mode)
193 THz (1550 nm) — Low λ ← Raman shift direction 210 THz (1427 nm) — High ν
Propagation Delay Timeline — Direction from Time (TDE Mode)
t₀ — pulse launch ← fiber propagation delay → t₀ + L/v_fiber — last aperture fires
The Physics — Step by Step

Five stages that make
direction = frequency

1
A short optical pulse enters the nonlinear waveguide

A sub-picosecond to ~10 picosecond laser pulse is launched into a specialty optical fiber with engineered nonlinear dispersion properties. The pulse enters with a well-defined carrier frequency ν₀. At this moment it is an ordinary optical pulse — what happens next depends on the fiber's nonlinear coefficient γ and dispersion β₂.

ν₀ ≈ 193 THz (λ = 1550 nm) — pulse enters nonlinear fiber
2
Soliton self-frequency shift — stimulated Raman scattering

The pulse's high peak intensity excites stimulated Raman scattering. Photons transfer energy to molecular vibrations in the glass, generating new photons at lower frequencies in the trailing edge. The carrier frequency decreases monotonically as the pulse travels. This is the soliton self-frequency shift (SSFS) — a continuous, passive, physics-driven process.

dν/dz ≈ −γ · T_R · P² / |β₂| — Raman shift rate
3
Each aperture radiates at a unique carrier frequency

Apertures along the fiber (V-grooves, oblique gratings, exposed cladding) radiate light outward at the frequency the pulse has reached at that position. Aperture #1 radiates blue (high ν). Aperture #12 radiates red (low ν). Each direction carries a unique spectral fingerprint — encoded by position, readable by frequency.

ν(z) = ν₀ − α·z  — frequency at aperture position z
4
Coherent receiver identifies direction from return frequency

Return pulses are measured by a coherent receiver using a local oscillator. The beat frequency identifies which aperture caused the return — and therefore which direction. No time-multiplexing, no channel selection electronics. Direction is read from spectral analysis alone at a single detector.

aperture_id = f(ν_return) → θ(aperture_id)
5
Range + Direction + Velocity from one coherent pulse

Range from time of flight. Direction from return frequency. Velocity from Doppler shift of the return relative to the expected aperture frequency. All three quantities from a single coherent measurement at one detector. This is the complete CFSRP information chain.

R = c·Δt/2  |  θ = θ(ν_return)  |  v = Δν_Doppler·λ/2
Side-by-Side Comparison

CFSRP vs. conventional LiDAR —
one guided medium (CPSI) vs. N parallel chains

The left simulation shows conventional multichannel LiDAR: N independent laser sources, each with its own driver, its own detector, its own signal chain — all firing simultaneously. The right simulation shows the CFSRP architecture: one pulse, one fiber, all directions encoded in frequency.

Conventional Multichannel LiDAR
Ψ SciPhAI — CPSI · One Guided Medium · TDE Primary · CFSRP Secondary
Conventional LiDAR
N laser sources — one per angular channel
N detectors — hardware cost scales with resolution
Electronic timing triggers with inherent jitter
No signal coding — crosstalk between vehicles
Direction determined by which channel fired
Velocity requires separate Doppler processing
Moving parts required for full FOV (spinning)
Cannot be embedded in vehicle glazing
Ψ SciPhAI CFSRP
One pulse source — one waveguide, all directions
One detector — hardware cost constant with resolution
Propagation physics — timing deterministic, zero jitter
Coded pulses — mathematical crosstalk immunity
Direction determined by return frequency alone
Velocity from Doppler shift of frequency fingerprint
No moving parts — propagation physics drives steering
Waveguide embeddable in vehicle window glass
Dual Operating Modes — Comparison

Same hardware.
Two distinct physics chains.

Both modes operate on identical CPSI hardware. The operating mode is selected by pulse duration, fiber operating regime (linear vs. nonlinear), and receiver architecture. A single SciPhAI platform can switch between modes or combine them depending on the application.

Property ⏱ TDE — Temporal Direction Encoding · Primary Mode 〜 CFSRP — Continuous Frequency-Shifting Radiating Pulse · Secondary Mode
Direction encoding principle Propagation delay — t_fire(i) = t₀ + zᵢ/v_fiber Carrier frequency — ν(z) = ν₀ − α·z
Fiber operating regime Linear — no nonlinear effects required Nonlinear — Raman/soliton shift required
Pulse duration Nanoseconds — relaxed constraint 1–10 ps — tighter, Fourier-limited
Detection hardware Direct detection — simple photodetector + timer Coherent detection — local oscillator + balanced receiver
Receiver complexity Very low — timing circuit only Higher — coherent receiver + DSP
Per-point velocity Not directly — separate measurement needed Yes — Doppler shift of frequency fingerprint
Hardware cost Lower — direct detection is simpler Moderate — coherent receiver adds cost
Timing jitter Zero — from propagation physics (both modes) Zero — from propagation physics (both modes)
Single detector for full FOV Yes — one detector regardless of resolution Yes — one detector regardless of resolution
No moving parts Yes — propagation physics drives steering Yes — propagation physics drives steering
Wavelength flexibility Broad — any wavelength the fiber supports Constrained by Raman shift spectral range
Atmospheric robustness High — direct detection insensitive to phase noise Good — coherent detection has some phase sensitivity
Best application fit All deployments — TDE is the default operating mode. Simpler hardware, lower cost, direct detection, broadest IP claim. Long range, dense traffic, velocity measurement needed
IP claim basis Core CPSI architecture — broadest claim CFSRP modality — unique, zero competitors
Strategic IP Note

The TDE (Temporal Direction Encoding) mode establishes the broadest independent claim: direction identification from propagation delay in a CPSI system requires neither nonlinear fiber nor coherent detection. Any competitor building CPSI-based illumination encounters this claim before reaching CFSRP. The two modes together create a layered IP structure — a foundational claim (TDE) with a uniquely novel enhancement on top (CFSRP).

Simulation Note

This simulation illustrates the CFSRP operating principle using physically motivated parameters. The frequency shift rate, fiber length, aperture positions, and pulse envelope are scaled for visual clarity. In a real CFSRP system, the soliton self-frequency shift rate depends on the fiber's nonlinear coefficient γ, the group-velocity dispersion β₂, the Raman response time T_R, and the pulse peak power P. Typical parameters: γ ≈ 10 W⁻¹km⁻¹ for highly nonlinear fiber, |β₂| ≈ 1 ps²/m, T_R ≈ 3 fs, P ≈ 1–10 kW peak. The total frequency shift across a practical fiber coil of 1–10 m can reach 10–50 THz — sufficient to resolve many hundreds of independent angular positions using wavelength-selective detection. The simulation color-encodes frequency using a continuous spectrum from violet-blue (high ν, short λ) through green to deep red (low ν, long λ), corresponding to the actual spectral shift direction of stimulated Raman scattering in silica fiber.

Source: US 2026/0056318 A1 ¶[0035], ¶[0057], ¶[0209]–[0216] · Inventor: Hamid Hatami-Hanza · Priority: October 17, 2024