Tesla’s Full Self-Driving (FSD) software was recently put to the test at the Commercial Club in Albury, Australia, which features one of the country’s largest installations of Tesla Superchargers. This testing aimed to see if the FSD could successfully navigate a multi-story carpark, a situation that many electric vehicle owners encounter regularly. The challenge included not only finding the charging station but also handling the steep ramps characteristic of such facilities.
Upon entering the destination for the Albury Supercharger, the Tesla attempted to navigate to the Commercial Club. However, instead of entering the carpark on the ground floor, it parked on the street adjacent to the building. This initial move indicated that the FSD was not prepared for the complexities of navigating multi-story carparks. The driver then took control and drove into the carpark, at which point FSD was reactivated to assess its capabilities.
Once back inside the carpark, the software attempted to exit rather than proceed to the charger. The driver re-entered the carpark, positioning the vehicle close to the on-ramp for the upper levels. When FSD was engaged again, the vehicle navigated up the ramp, pausing briefly at the apex due to a lack of visibility. As the incline obscured the view ahead, the vehicle hesitated, indicating a lack of confidence in the situation.
The driver pressed the accelerator, allowing the vehicle to move forward and gain the necessary visibility to reach the first level where the Supercharger bays are located. Interestingly, instead of stopping at a charging station, the Tesla continued navigating to the top level of the carpark. Upon reaching the coordinates of the Supercharger, the navigation system cleared the destination, which resulted in the vehicle entering a free-roaming mode instead of attempting to park.
After charging, the driver decided to test the FSD’s ability to navigate back down. Setting the destination for the ground level, the vehicle successfully maneuvered through the turns and down the exit ramp. However, the FSD guessed incorrectly at one point, turning left instead of right, which could have led to confusion in a busier setting. Fortunately, the area was clear of traffic, which mitigated potential issues.
Reflecting on the performance, it is evident that the vehicle’s ability to detect other vehicles, pedestrians, and obstacles remains strong. The primary area for improvement lies in its confidence while navigating situations with limited visibility, such as the top of the ramp. It may also be beneficial for the FSD to incorporate three-dimensional navigation capabilities to better understand the vertical components of carpark navigation.
When the destination for the Albury Supercharger was set, the navigation system provided guidance but appeared to only consider two-dimensional positioning. This limitation may have contributed to the vehicle’s failure to seek out the Supercharger bays after reaching the designated floor.
In conclusion, while Tesla’s FSD demonstrates significant advancements in navigating complex environments, there remains room for refinement, particularly in multi-story settings. Observing how the software adapts to various scenarios will be crucial as Tesla continues to develop its autonomous driving technology.
