High-Uptime Sensor Streaming

A lightweight sensor uptime utility designed for robots and autonomous vehicles that rely on continuous sensor data for safe operation.

Instant-On
Continuous Streaming

Deploy on
1-100,000 Systems

Zero-Copy for No Latency Impact

Maintain Stability During Sensor Failures

Performance Tracking for Root Cause Analysis

Automatic Restart Attempts

Perception Sensor Fusion & Calibration

Plug & Play Perception

Don’t waste your time with flimsy firmware and sparse SDKs. Our streaming tool manages drivers, firmware, and data flow to keep your perception system running smoothly. It proactively manages sensor uptime with runtime tools and provides early awareness of pending failures with performance checks and proactive notifications.

Quickly test different sensors & modes

Engineered in Rust for safety & speed

Compatible with most operating systems

Work with your existing communication framework: ROS, protobufs, files

Nonstop Sensor Stability

When put to the test, our streaming tool increased the mean time between resets by 99 percent — and when sensors do fail, we can get them back online in under 10 seconds. Eliminating sensor failures will save months of engineering effort, reduce maintenance costs, and help keep your customers happy.

Frequently Asked Questions

How does the streaming tool work?

Our streaming software is deployed to your device host in a Docker container. Our software monitors the streams coming from your sensors and identifies when a sensor is entering a potential failure mode. If a failure mode is detected, our streaming tool performs an automatic restart for the failing sensor, and can send an optional user alert.

Why use it?

Our streaming tool eliminates hacks like separate boards to power cycle sensors. Its zero-copy, containerized design installs easily with minimal impact on system memory and latency. Our software preserves overall system stability in the case of a sensor failure. In timed tests, our tool brought failed sensors back online in under 10 seconds.

Which sensors can it manage?

Our streaming tool is available now for all Intel RealSense™ D-series sensors and will soon support additional depth sensors, cameras, LiDARs and radars.

How much does it cost?

It's included in our platform! Contact sales@tangramvision.com for more info or sign up for a trial of our perception platform.

Tangram Vision

Time To Deploy

One Hour

Modalities Supported

Cameras, Depth (Now); LiDAR, IMU, Radar (Soon)

Multi Modality Process

One Dataset Calibrates All Sensors

Quality Checks

Target Correction, Error Covariance

Data Management

Dashboard or API

Deployment

Cloud, Local Server, Edge

Support

Forum and Email with SLAs

Open Source

Time To Deploy

Weeks to Months

Modalities Supported

Cameras, Depth

Multi Modality Process

Separate, Sequential Datasets

Quality Checks

None

Data Management

None

Deployment

Local Server

Support

Community

Build It Yourself

Time To Deploy

Months to Years

Modalities Supported

Cameras, Depth, ?

Multi Modality Process

Separate, Sequential Datasets

Quality Checks

None

Data Management

None

Deployment

Local Server

Support

None

Perception Resources

Depth Sensor Visualizer

Go >

Calibration Target Library

Go >

Sensor Roundup Series

Go >

Build Calibration For Prototypes