Volvo, VERSES release study on algorithm to predict what’s behind obscured objects
By onAnnouncements | Technology
Volvo and VERSES research teams recently published a paper using an algorithm to predict the appearance of pedestrians, cyclists, and cars that are obscured behind stationary vehicles and objects, according to a press release.
The release claims the advancement is “beyond current capabilities of autonomous vehicle AI.”
VERSUS, a cognitive computing company specializing in intelligent systems, created the algorithm used in the Genius Beta Partner collaboration.
“As the automotive industry progresses towards fully autonomous self-driving cars, predicting where unseen obstacles like people or bicyclists may be or which trajectory they may be on has been a significant unsolved safety challenge,” said Gabriel René, VERSES CEO, in the release. “We believe the inability of current autonomous driving systems to overcome this hurdle is holding back the AV industry worldwide. Volvo Cars is globally recognized for its unwavering commitment to vehicle safety. So, they were the perfect partner to work with to showcase how VERSES can help solve this problem.
“We believe this research project with Volvo Cars, part of our Genius Beta project, demonstrates a major advancement in autonomous vehicle safety capability. We expect the research project to pave the way for safer streets for pedestrians, cyclists, cars, robots, and beyond.”
The paper, “Navigation under uncertainty: trajectory prediction and occlusion reasoning with switching dynamical systems,” covers completed experiments illustrating capabilities using the Waymo open dataset, the release says.
“Prior works typically neglect to maintain uncertainty about occluded objects and only predict trajectories of observed objects using high-capacity models such as transformers trained on large datasets,” the paper says. “While these approaches are effective in standard scenarios, they can struggle to generalize to the long-tail, safety-critical scenarios. In this work, we explore a conceptual framework unifying trajectory prediction and occlusion reasoning under the same class of structured probabilistic generative model, namely, switching dynamical systems.”
Transformer architecture uses high-capacity function approximators for trajectory prediction, the paper says. It says the state-of-the-art (SOTA) models typically function with the assumption that all objects in the scene are fully observed.
Switching dynamical systems divides the modeling of complex continuous dynamics into a finite set/mixture of simple dynamics arbitrated by switching variables, according to the paper.
“The main attractiveness of this class of models is that it provides a unified representation where hierarchical compositions generalize to both prototype-based trajectory prediction and object-centric occlusion reasoning,” the paper says.
Future work will incorporate auxiliary information, such as road graphs, to improve prediction accuracy and implement efficient inference algorithms that are particularly suitable to this family of models, according to the paper.
A recently updated position statement from Volvo on scanning and diagnostics says that the OEM’s latest technology incorporates numerous control modules, sensors, and camera systems that support various vehicle functions.
“These components are an integral part of a vehicle’s operational systems. In a collision, these components could be damaged in ways that are not visually evident, so they must be inspected and evaluated, regardless of whether the vehicle displays a failure via the instrument cluster; this will help ensure that any possible lack of functionality will be discovered and resolved.”
Pre-repair scanning is a crucial step that is required prior to performing any and all work on the vehicle, the statement says. Post-repair scanning and diagnosis of the vehicle is required for any work that is done on the vehicle.
“This ensures the vehicle’s safety and autonomous systems are functioning correctly,” the statement says. “Any safety or autonomous systems activated during a collision may require further inspection, initialization, calibration, or replacement demonstrating your diligence in the repair process.”
Images
Feature photo courtesy of jetcityimage/iStock