E. Jared Shamwell
DEVCOM Army Research Laboratory, University of Maryland, College Park
Papers
3
Total Citations
12
H-Index
3
About
E. Jared Shamwell is a researcher advancing the frontiers of autonomous robotics through bio-inspired sensor fusion and deep learning. His work focuses on enabling robust state estimation for size, weight, and power (SWaP)-constrained robotic systems, a critical challenge for field-deployable robots. Shamwell’s major contributions center on developing unsupervised deep neural networks that intelligently fuse heterogeneous sensor data—such as vision and motion estimates—to overcome the limitations of low-cost, low-power hardware. His most cited paper (2017, 5 citations) introduces a deep neural network approach for fusing vision with heteroscedastic motion estimates, directly addressing the computational bottlenecks in low-SWaP platforms. A second influential work (2018, 4 citations) presents the Multi-Hypothesis DeepEfference (MHDE) network, an unsupervised convolutional-deconvolutional architecture that learns to combine noisy sensor streams for improved visual motion estimation. Notably, Shamwell’s research draws inspiration from biological vision, as seen in his “DeepEfference” concept (2017, 3 citations), which models how the human brain maintains visual constancy during saccades—a principle applied to robotic localization. By bridging neuroscience and engineering, Shamwell’s work offers practical pathways for more capable, autonomous robots operating under severe resource constraints.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3