About NovoLux & Manifold

NovoLux puts the powerful health and performance benefits of natural light virtually anywhere. Their product makes it possible to replicate natural light and a real-time view of the outdoors in any windowless, sunless space.

Manifold is a full-service AI consulting company offering a complete range of engineering services, including machine learning, data science, data engineering, devops, cloud, and edge.

Challenge

NovoLux had developed unique core technology for intelligent lighting. They approached Manifold to accelerate product development.

“The Manifold team is sharp and accessible. In 12 weeks, Manifold’s process transformed our AI goals into a lean prototype and a large product opportunity. We’re now working with them to realize it.”

— Pooja Devendran, NovoLux CEO

The highest priority was to streamline and improve the prototype. Our client’s prototypes used off-the-shelf components that required a labor-intensive and manual installation process. Deploying these prototypes to their customers required frequent and intense recalibration of sensors and a resource-intensive setup process routing cables through a customer’s building. There was no way to remotely monitor deployments, so NovoLux’s engineers had to drive out to implementation sites to provide customer support in person. In short, our client’s ability to scale their deployments and service offerings was limited.

Solution

Manifold took a prototype implementation and transformed it into a shippable, scalable, upgradable product. By turning hardware problems into software solutions, we were able to remove several components from their system. NovoLux now has blueprints for a remotely configurable, scalable IoT platform, and a product roadmap they can implement on that platform.

Product Development

Manifold took a lean approach to product development, which started by detailing user stories of our client’s enterprise customers, employees, and IT department. We were able to capture the existing pain points related to manufacturing, installation, and customer support, and understand how to improve the overall workflow. We then researched similar existing products and solutions and created a plan to upgrade components in the right sequence.

Hardware Optimization

Manifold selected and tested various hardware configurations for controlling the ambient light output, including multiple LED sources. We settled on hardware that could interface with any dimmer on multiple lighting products. We guided design to enable NovoLux to make quick, iterative decisions about product features while also reducing bill of materials (BOM) cost and improving ease of deployment and manufacture.

Wherever possible, we chose off-the-shelf components over custom hardware development, and turned hardware-specific implementations into software implementations. We improved the client’s video-routing and hardware infrastructure with custom, upgradable software. We replaced their recording hardware with a wireless solution that enabled remote administration.

After experimenting with TV OS APIs and edge computers, we selected a system that could scale in memory, storage, and compute power and would be future-proof for feature expansion.

Computer Vision

Manifold’s expertise with computer vision enabled us to develop production-grade computer vision software to pipe real-time video to the screen, as well as parse video images, make calculations, and control LED lighting. We created a custom proprietary algorithm that accounted for time, location, environmental activity, and incorporated a frame-rate that would allow for perceptible changes.

Remote Administration

The above parameters can now easily be configured from a settings file, and more sophisticated algorithms can be implemented in the field using a remote software upgrade. Furthermore, we allowed for a remote administration system that is both secure and scalable.

NovoLux can now securely log in to a single portal and remotely monitor all of their deployed IoT devices. They can remotely view log files, get notifications when installations are not behaving properly, configure site-specific settings file, and perform remote software upgrades as more features are added to the software.

Robust Design

By limiting local user access to the system, and creating fail-safe software, the IoT system we created is robust to network and power failure, plugging/unplugging, and physical reset. This “just-works” design cuts down on the amount of customer support required.