An Apple Patent reveals a method for an autonomous vehicle to be able to merge or make a lane change safely
In late March the US Patent & Trademark Office published a Project Titan related patent application from Apple that describes systems and methods for calculating how and when an autonomous vehicle could safely merge into traffic or make lane changes safely.
In Apple's patent background they state that navigating a vehicle from a first step to a second step along a route often includes directing the vehicle into a flow of traffic involving multiple vehicles, in other words, merging.
Entering a flow of traffic generally involves avoiding conflict with other vehicles by directing the vehicle into to a gap between vehicles. Such gaps, however, are predicated on whether a corresponding vehicle will yield to allow the vehicle to enter the flow of traffic, and it may be challenging to determine whether the corresponding vehicle intends to yield.
These challenges involving uncertainty in whether a vehicle will yield are exacerbated in the context of autonomous vehicles, as the uncertainty is often conventionally resolved through behavioral cues or exchanges between the operators of vehicles. It is with these observations in mind, among others, that various aspects of the present disclosure were conceived and developed.
Apple's invention covers technology related to systems and methods for directing an autonomous vehicle into a flow of traffic given an uncertainty in yielding agents at an interaction zone.
Generally, as the autonomous vehicle navigates towards an interaction zone in a flow of traffic, each vehicle in the flow of traffic is designated as a non-yielding agent or a yielding agent and gaps in the flow of traffic between these agents are identified.
At a planning cycle, a motion plan is generated to enter the flow of traffic at a first gap between a non-yielding agent and a yielding agent. If it is determined that the first gap exceeds an initial confidence threshold, the autonomous vehicle enters the interaction zone at the first gap.
If the first gap is uncertain where there is uncertainty in whether the yielding agent will yield to the autonomous vehicle at the interaction zone, the autonomous vehicle hedges to enter the flow of traffic at a second gap following the yielding agent without computing a different motion plan.
The example implementations discussed in Apple's patent reference a traffic environment involving a first traffic lane merging into a second traffic lane, such as in the context of a highway onramp.
However, it will be appreciated by those skilled in the art that the presently disclosed technology is application in other traffic environments involving interactions among vehicles, including, without limitation, lane merges, lane changes, intersections, parking lots, and/or other shared spaces.
Apple's patent FIG. 1 is a block diagram of an example traffic environment with an autonomous vehicle hedging for different gaps in a flow of traffic; FIG. 2 shows a block diagram of the example traffic environment where the autonomous vehicle entered a first gap in the flow of traffic.
Apple's patent FIG. 4 above is a block diagram of the example traffic environment where the autonomous vehicle hedges to stop.
Apple's patent FIG. 9 below shows an example vehicle control system for an autonomous vehicle; FIG. 10 illustrates example operations for controlling an autonomous vehicle; FIG. 11 illustrates example operations for controlling an autonomous vehicle.
Apple's patent FIG. 12 above depicts a block diagram of an electronic device including operational units arranged to perform various operations of the presently disclosed technology; FIG. 13 illustrates example operations for controlling an autonomous vehicle.
Further to FIG. 9 above, Apple shows an example vehicle control system (#600) for the autonomous vehicle. In one implementation, the vehicle control system includes a perception system #602, a motion control system #604, and vehicle subsystems #606.
The perception system includes one or more sensors, such as imagers, LIDAR, RADAR, etc., to capture information regarding objects in a field of view of the autonomous vehicle.
For example, the perception system may capture traffic flow data for the flow of traffic, including a location of and motion information regarding the vehicles within the flow of traffic, such as the vehicles 110-114 of FIG.1. The perception system may further capture information to define an interaction zone, as well as measuring, tracking, and/or estimating an actual distance of the autonomous vehicle #102 to the interaction zone.
For finer details, review Apple's patent application number 20210089041 that was published on March 25 by the U.S. Patent Office. The patent was originally filed in Q3 2020.
Considering that this is a patent application, the timing of such a product to market is unknown at this time.