Hoodtronics
HOODTRONICSAI

Research Manufacturing Solutions

From embedded firmware and 3D manufacturing to patent-pending image codecs and neural reconstruction.

NSF SBIR Phase I TRL 4 · Lab Validated USPTO Provisional Filed TechConnect 2026
Hoodtron piloting a drone
124.5:1Binary Compression
62.3:1On-Air Compression
42.6 dBPSNR Quality
AES-256GCM Encrypted
5 yrR&D Depth
ImageGenAI Platform

Three-Layer Visual Intelligence Pipeline

Patent-pending architecture. Each layer is independently deployable. The full pipeline captures, compresses, and reconstructs images across a LoRa radio link.

Hoodtronics racing vehicle
01 Capture

CVisionBot

ML-gated 2-DOF pan/tilt ESP32-CAM robot. On-device inference filters frames before transmission. Only high-confidence detections enter the codec pipeline.

02 Codec

CACADU

7-stage CPU codec: linear encoding → AES-256-GCM → FEC → LoRa radio. 124.5:1 binary compression. Recovers frames under packet loss conditions.

03 Reconstruct

ImageGenAI CNN

CPU quality validation feeds a GPU generative reconstruction network. Restores full-fidelity images from compressed LoRa transmissions via autoencoder.

Repository
trueverbals4 / ImageGenAI
View on GitHub →
Hoodtron — Autoencoder Analysis
ImageGenAI CNN — Architecture

Convolutional Autoencoder

The reconstruction network learns a compressed latent representation. The encoder maps degraded frames to a bottleneck; the decoder restores full fidelity.

INPUT LoRa RX CONV1 64 filters CONV2 128 CONV3 256 ENCODER LATENT SPACE BOTTLENECK compressed vector TCONV3 256 TCONV2 128 TCONV1 64 filters DECODER OUTPUT Restored MSE LOSS · BACKPROP GPU Training & Inference CPU Quality validation before GPU pass
Repository
trueverbals4 / ImageGenAI
View on GitHub →
Products & Services

What We Build

Research-grade hardware and software. Available for licensing, collaboration, and sub-award research partnerships. All technical documentation on GitHub.

Hoodtron — Hoodtronics manufacturing mascot
📡
Edge Hardware

CVisionBot

Open-source ESP32-CAM pan/tilt robot. ML inference on-device. MJPEG stream compatible with any Python/OpenCV host. PETG 3D-printable enclosure. 10 STL files included.

View on GitHub
🔐
Embedded Codec

CACADU

CPU-side 7-stage image codec. AES-256-GCM encryption. FEC error recovery. LoRa radio integration. 124.5:1 binary compression. 42.6 dB PSNR. Provisional patent filed Oct 2026.

View on GitHub
🧠
AI Platform

ImageGenAI

End-to-end visual intelligence platform. CVisionBot capture → CACADU compression → CNN reconstruction. NSF SBIR Phase I applicant. TRL 4 — validated in laboratory.

TechConnect 2026 →
🖨️
Additive Mfg

3D Printing

Custom enclosures, brackets, and product housings for embedded electronics. CAD-to-part via Fusion 360 and SolidWorks. FDM and resin. PETG, PLA, and ABS material options.

Request a Quote →
🧪
Prototyping

Silicone Molding & Casting

Low-volume silicone molds and urethane casts for custom parts and prototypes. Suitable for STEM kits, sensor housings, and embedded device enclosures requiring irregular geometry.

Request a Quote →
🛡️
IP Service

Customer-Owned IP Device

You own the design. Hoodtronics builds it. We manufacture embedded devices to your specification — firmware, PCB, enclosure — and transfer full IP. No licensing. No royalties. Your product.

Start a Project →
CVisionBot — Live Interface

Pan/Tilt Cam Web Control

ESP32-CAM robot — MJPEG stream with 2-DOF pan/tilt control over LAN. Enter your device IP below to connect.

No Stream — Enter IP to Connect
Offline
Device IP
Pan / Tilt
Telemetry
PAN 90°
TILT 90°
FPS --
Repository
trueverbals4 / VisionBot
View on GitHub →
CACADU — 7-Stage CPU Codec

Compression Pipeline in Action

Every frame passes through seven deterministic CPU stages before LoRa transmission. No GPU required — the entire codec runs on embedded ARM.

Hoodtronics CACADU vehicle
cacadu — 7-stage codec pipeline
hoodtronics@cacadu-lab:~$ ./cacadu encode --input frame_001.raw --key firmware_key.pem
╔══════════════════════════════════════════════════════╗
║ CACADU v1.0 — Layer II Visual Codec ║
║ Hoodtronics — Indianapolis, IN ║
╚══════════════════════════════════════════════════════╝
Stage 1 › Linear Encoding ....................... ✓ 0.3 ms
Stage 2 › AES-256-GCM Encrypt .................. ✓ 1.2 ms
Stage 3 › Reed-Solomon FEC ..................... ✓ 0.8 ms
Stage 4 › Frame Packetize ...................... ✓ 0.1 ms
Stage 5 › LoRa TX Queue ......................... ✓ 0.0 ms
Stage 6 › LoRa Radio Transmit .................. ✓ 2.1 s
Stage 7 › RX Reconstruct → ImageGenAI CNN ....... ✓ ongoing
Input : 320×240 px → 76,800 bytes
Output : 616 bytes (binary compressed)
Ratio : 124.5:1 binary · 62.3:1 on-air
PSNR : 42.6 dB (post-CNN reconstruction)
Crypto : AES-256-GCM · PLACEHOLDER KEY — lab only
[CACADU] Transmitting via LoRa @ 915 MHz SF10 BW125...
83% · ETA 2.1s
hoodtronics@cacadu-lab:~$
IP Protected — Patent Pending
trueverbals4 / cacadu-codec
CACADU — Radio Link

LoRa Calibration & Link Profile

CACADU transmits compressed frames over LoRa 915 MHz. The chirp spread-spectrum waveform survives low SNR and long range. Calibration sets spreading factor, bandwidth, and coding rate.

Hoodtron — LoRa calibration
Radio Config
Frequency
915 MHz
Band
US ISM
Spreading Factor
SF 10
Bandwidth
125 kHz
Coding Rate
4/5
TX Power
20 dBm
RSSI
−87 dBm
SNR
+6.5 dB
Transmission Rate
183 b/s · 62.3:1 on-air
✓ PASS
0 −20 −40 −60 −80 902 908 915 922 928 Frequency (MHz) noise 915 MHz BW 125 kHz upchirp sweep → LINK BUDGET RSSI: −87 dBm SNR: +6.5 dB SF10 CR4/5 LOCK
IP Protected — Patent Pending
trueverbals4 / cacadu-codec
Research Recognition

Credentials & Validation

Hoodtronics research is verified, protected, and recognized at the federal and industry level.

NSF · Federal Grant

SBIR Phase I Applicant

Submitted to the National Science Foundation Small Business Innovation Research program. UEI: Z3TLHYQ1QMT3. Active SAM.gov registration. Funding targets ImageGenAI commercialization.

TechConnect World · 2026

Innovation Showcase

ImageGenAI accepted to TechConnect World Innovation Conference 2026, Raleigh, NC — March 10–12, 2026. Technology Readiness Level 4. Primary: AI & Digital Transformation.

USPTO · IP Protection

Provisional Patent Filed

Provisional patent filed October 3, 2026 covering the CACADU codec and ImageGenAI visual intelligence pipeline. Non-provisional filing due before October 3, 2027.

The Face of Hoodtronics

Meet Hoodtron

Hoodtron is our research manufacturing mascot. Built in the lab, hood up, soldering iron in hand — representing the hands-on work behind every Hoodtronics system.

Every board we fabricate, every enclosure we print, every firmware line we flash — that's Hoodtron at work. Research manufacturing isn't a tagline. It's what we do every day.

3D Printed Enclosures PCB Fabrication Embedded Firmware LoRa Radio Systems
Hoodtron — the face of Hoodtronics
INNOVATE · ENGINEER · ELEVATE HT-01
About Hoodtronics

Hoodtronics Research Manufacturing

Hoodtronics is a research manufacturing company. We design, build, and commercialize embedded AI systems — and deliver AI solutions to clients through HoodtronEI. Our work spans embedded firmware, machine learning, 3D printing, and PCB fabrication.

As part of our long-term community mission, Hoodtronics commits 10% of future sales toward placing 100 AI Vision Bots in children's STEM programs by August 2026.

Founded 2020. Two of four designed products commercialized. Continuous R&D output across embedded AI and vision systems.

3D Printing Services. Custom enclosures, brackets, and STEM product housings. We design and fabricate parts optimized for embedded electronics using CAD and additive manufacturing.

AI Solutions via HoodtronEI. Credentialed access to our codec API, CNN model weights, and LoRa telemetry stack. Built for AI protocol engineers. Adversarial cloud server platform coming soon.

Federal Research Pipeline. Active SAM.gov registration, NSF SBIR Phase I application, and Research.gov enrollment under UEI Z3TLHYQ1QMT3.

University & Lab Partnerships. Open to sub-award research collaborations and federal co-PI arrangements. Contact us to discuss scope.

⬡

HoodtronEI AI Solutions

HoodtronEI connects your AI use case to the right protocol. Tell us what you need — we map it to a solution. Credentialed engineers get direct platform access.

◈ Tell Us Your AI Use Case

Describe the problem you want AI to solve. HoodtronEI maps your use case to a solution we can implement — image transmission, edge inference, computer vision, or a custom pipeline.

Describe your use case · Hoodtronics maps you to a solution
🔐
Engineer Access
CREDENTIALED ONLY

AI protocol engineers with verified credentials get direct access to the full HoodtronEI platform stack.

Codec API
CACADU encode/decode
T1
Model Weights
CNN autoencoder checkpoints
T2
LoRa Telemetry
Live radio stream
T2
CVisionBot SDK
Full hardware control
T3
Requirements: AI protocol engineering background. NSF · DoD · University affiliation accepted at T1. NDA required for all tiers.
Private Repository
trueverbals4 / HoodtronEI  ·  Private
Legal

Privacy Policy

Effective Date: November 1, 2026  |  Hoodtronics LLC, Indianapolis, IN

1. Information We Collect

We collect only the information you voluntarily provide. This includes your name and email address when you submit a contact or access request form. We do not use tracking pixels, behavioral analytics, or third-party ad networks on this site.

2. How We Use Your Information

Contact information is used solely to respond to your inquiry or to evaluate a research collaboration or licensing request. We do not sell, rent, or share your personal data with any third party for marketing purposes.

3. Research Data

The HoodtronEI Engineering Survey collects de-identified workflow and prompt-engineering responses for academic research purposes. No personally identifiable information is published in research outputs. Survey participation is voluntary and participants may withdraw at any time.

4. Cookies & Storage

This site does not use cookies. No data is stored on your device beyond standard browser session memory required for page functionality.

5. Intellectual Property

All content on this site — including the CACADU codec architecture, LoRa-ImageGenAI pipeline, and associated firmware — is proprietary to Hoodtronics LLC and protected by a USPTO provisional patent filed October 3, 2026. Unauthorized reproduction or reverse engineering is prohibited.

6. Data Retention

Contact inquiry emails are retained for up to 24 months for correspondence records. You may request deletion of your data at any time by emailing trueverbals@hoodtronics.com.

7. Contact

Questions about this policy may be directed to:
Hoodtronics LLC — Indianapolis, IN
3410 N High School Rd, Suite G, Indianapolis, IN
trueverbals@hoodtronics.com

Hoodtron East
Collaborate

Partner With Hoodtronics

We are seeking university partners, federal lab collaborators, and licensing inquiries for the ImageGenAI platform.

Hoodtron West
Hoodtronics Portal

Account Access

Anyone can submit the access survey. HoodtronEI professional tool access requires credentials.

Forgot password?
🔬
RESEARCHER
API & data access
🤝
PARTNER
Collab tools
🏭
ENTERPRISE
Full suite