Projects

At Olson Scientific our research and development projects have spanned a wide range of industries and focuses, but are all centered around the core critical chain technologies common to our areas of focus: Aerospace, Autonomy, and AI.

A sample of these projects follow, to illustrate the type of technologies we frequently work with.

Simulations + VR

We frequently create simulations and interactive virtual environments for training, testing, virtual user manuals, games, and other purposes.

Hybrid Mobility Robot

This PX4 and ROS based hybrid modality UAS was created as a pre-production prototype for a client to use in fundraising It utilizes edge computing to enable advanced autonomy features and can both roll and fly.

SOAR

SOAR is a PX4 based sailplane that uses thrust vectored coax motors to self launch and seek thermals. SOAR has an onboard AI system that controls the thermal seeking while maintaining communication with a ground station that can relay it additional information, including premade thermal maps and sensor data.

ROLLFLY / ROLLFLYGLIDE

Utilizing active thrust vectoring and custom dynamic control code we were able to take hybrid mobility to the next level – averaging efficiency gains of 10x over traditional quadcopters and 2x over the HMR.

By switching to a larger form factor and using an adaptive geared power train we can adapt the rollfly drone to smooth controlled movement in water air or on land.

WIRE / DEEPSCAN

One of our primary focuses is non-invasive neural human machine interfacing for virtual reality and robotic systems control.

EEG and EMG and other real time non invasive monitoring methods allow for deep understanding of the user’s body when utilizing custom AI models trained to understand the signals. This allows for real time interpretation and prediction of the user’s desired control input, emotional state, and a large number of other factors.

Project eXP

VR motion sickness prevention via noninvasive galvanic vestibular stimulation. Initial tests indicated similar methods are likely to also have promise for tactile proprioceptor reinforcement in simulations.

SEED

Distributed edge processing sensor network with semantic-rich real-time centralized AI processing and intelligence.

Unlike NVIRs or AI cams, the SEED system generates a deep understanding of the area of interest – allowing it to autonomously take actions.

Multimodal RAG, Custom Agents, etc

Combining vectored and other knowledge storage with our custom LLM based reasoning agents using structured logic and semantically enabled processing have demonstrated highly effective results.

Using these tools we have produced bespoke search engines, virtual semantically aware agents, and other similar systems that allow for autonomous semantic based packaging and processing of unstructured data, real time generation of VR and 3D environments, and even AI based Aerospace design tools.

*Please note, not all projects and details can be shown due to confidentiality, privacy, national security and legal agreement constraints. Please check back soon as we are still expanding our project list!