ARKANSAS, Nov 20 (Future Headlines)- Tesla has initiated the rollout of a new Full Self-Driving (FSD) Beta software update, marked as version 11.4.8, introducing a series of improvements as the company continues its quest to enhance autonomous driving capabilities. This update precedes the highly anticipated v12 update, which is expected to transition the Full Self-Driving system out of beta status. While Tesla’s FSD Beta enables autonomous driving, drivers are still required to remain vigilant and ready to take control when necessary.
Simplified Activation and Disengagement:
Change: Added an option to activate Autopilot with a single stalk depression instead of two, simplifying the activation and disengagement process for drivers.
Efficient Video Module for Vehicle Detection:
Change: Introduced a new efficient video module to the vehicle detection semantic, velocity, and attributes networks.
Benefits: Increased performance at lower latency, achieved through a multi-layered hierarchical video module that optimizes computation.
Enhanced Object Detection:
Change: Improved distant crossing object detections by an additional 6%.
Benefits: Improved precision of vehicle detection through autolabeling enhancements and the introduction of the new video module.
Improved Cut-in Vehicle Detection:
Change: Increased precision of cut-in vehicle detection by 15%.
Benefits: Utilized additional data and video architecture changes for improved performance and latency.
Reduced Velocity and Acceleration Errors:
Change: Reduced vehicle velocity error by 3% and vehicle acceleration error by 10%.
Benefits: Improved autolabeled dataset, introduced the new video module, and aligned model training and inference more closely.
Lower Latency for Vehicle Semantics Network:
Change: Reduced the latency of the vehicle semantics network by 15%.
Benefits: Implemented the new video module architecture with no compromise to performance.
Enhanced Pedestrian and Bicycle Detection:
Change: Reduced the error of pedestrian and bicycle rotation by over 8%.
Benefits: Leveraged object kinematics more extensively in autolabeled datasets when optimizing pedestrian and bicycle tracks.
Improved Vision Park Assist Predictions:
Change: Improved geometric accuracy of Vision Park Assist predictions by 16%.
Benefits: Leveraged 10x more HW4 data, tripled resolution, and increased overall stability of measurements.
Increased Path Blockage Lane Change Accuracy:
Change: Improved path blockage lane change accuracy by 10%.
Benefits: Implemented updates to static object detection networks.
Context and Anticipation for v12:
Neural Net-Based Vehicle Controls: While the v11.4.8 update brings substantial improvements, the true game-changer is expected to come with the v12 update, which is anticipated to include neural net-based vehicle controls.
Elon Musk’s Full Self-Driving Promise: Tesla CEO Elon Musk has previously expressed the goal of achieving true Full Self-Driving capabilities, allowing drivers not to pay attention anymore. This milestone was initially projected to happen by the end of the year.
Tesla’s continuous efforts to enhance its Full Self-Driving system demonstrate the company’s commitment to pushing the boundaries of autonomous driving technology. The incremental improvements introduced in the v11.4.8 update pave the way for the upcoming v12 release, which is expected to be a significant milestone in Tesla’s journey toward achieving true Full Self-Driving capabilities. As Tesla remains at the forefront of autonomous vehicle development, drivers and enthusiasts eagerly anticipate the transformative features that v12 is likely to bring to the driving experience.
Reporting by Alireza Sabet; Editing by Sarah White