Event-based cameras are dynamic vision sensors that provide asynchronous measurements of changes in per-pixel brightness at a microsecond level. This makes them significantly faster than conventional frame-based cameras, and an appealing choice for high-speed robot navigation. While an interesting sensor modality, this asynchronously streamed event data poses a challenge for machine learning based computer vision techniques that are more suited for synchronous, frame-based data. In this paper, we present an event variational autoencoder through which compact representations can be learnt directly from asynchronous spatiotemporal event data. Furthermore, we show that such pretrained representations can be used for event-based reinforcement learning instead of end-to-end reward driven perception. We validate this framework of learning event-based visuomotor policies by applying it to an obstacle avoidance scenario in simulation. Compared to techniques that treat event data as images, we show that representations learnt from event streams result in faster policy training, adapt to different control capacities, and demonstrate a higher degree of robustness to environmental changes and sensor noise.

Flowchart-based programming languages are useful because they are easy to interpret on the user end, and have recently become more popular in implementation in certain fields. In this framework, a program is recognized as a flowchart. In this flowchart, different parts of the program can be executed concurrently or in an optimized order using scheduling algorithms. However, more complex equations, such as large heuristic algorithms for path finding, tend to result in huge slowdowns in execution time due to execution delays in the pipeline. The solution of this paper is to create an algorithm that uses multithreading to maximize the speedup of a given program, while taking into account both theoretical limits as well as implementation limits (due to thread creation and destruction, synchronization, and load balancing). This algorithm also incorporates a new form of branch prediction, which will limit the number of calculations and threads needed in these workflows. Implementation and analysis of the results of this algorithm's use are explored at the end of this paper.