PITA: Physics-Informed Trajectory Autoencoder
This addresses the need for generating edge case scenarios in safety-critical robotic applications, though it is incremental as it builds on existing autoencoder methods.
The paper tackles the problem of generating physically plausible trajectories for robotic system validation by proposing PITA, a Physics-Informed Trajectory Autoencoder that incorporates a physical dynamics model into the loss function, resulting in smooth trajectories that adhere to the model and outperform a normal autoencoder and a state-of-the-art action-space autoencoder on a real-world vehicle dataset.
Validating robotic systems in safety-critical appli-cations requires testing in many scenarios including rare edgecases that are unlikely to occur, requiring to complement real-world testing with testing in simulation. Generative models canbe used to augment real-world datasets with generated data toproduce edge case scenarios by sampling in a learned latentspace. Autoencoders can learn said latent representation for aspecific domain by learning to reconstruct the input data froma lower-dimensional intermediate representation. However, theresulting trajectories are not necessarily physically plausible, butinstead typically contain noise that is not present in the inputtrajectory. To resolve this issue, we propose the novel Physics-Informed Trajectory Autoencoder (PITA) architecture, whichincorporates a physical dynamics model into the loss functionof the autoencoder. This results in smooth trajectories that notonly reconstruct the input trajectory but also adhere to thephysical model. We evaluate PITA on a real-world dataset ofvehicle trajectories and compare its performance to a normalautoencoder and a state-of-the-art action-space autoencoder.