Change Lanes

Figures and examples
Changing lanes is an activity that helps illustrate how CAVs are different from cars with drivers. Three basic principles apply to changing lanes on a CAVWAY:
  1. A CAV may only change lanes into an open box, for example B, C or E in Figure 3 at right;
  2. To assure that a box will open up for a CAV needing to change lanes, CAVs in one lane (e.g., Lane 1 in Figure 3 at right) must be traveling at a different speed from CAVs in the adjacent lane (Lane 2); and
  3. To assure that there is such an open box in the target lane, the CAVWAY cannot be overloaded. CC will use a loading algorithm that assures that any CAV granted CAVWAY access will be able to exit at its destination.
Lane-change protocols
In the current example design (modeled under CSIM), CC determines lane-change protocols and broadcasts those to all CAVs. For example, on a 2-lane CAVWAY, CAVs exiting or moving from this CAVWAY to another at the next node must be in the number 2 (right, slow) lane. Other CAVs must be in the number 1 (left, fast) lane. Every CAV periodically assesses its situation, and then either stays in its current lane or changes lanes as indicated by the protocol.
A CAVWAY typically will consist of three types of “lane.” CAVs would use an Express Lane to bypass nodes, a Local Lane to enter traffic flow or exit at the next node, and an Access Lane to move between a node and a Local Lane. Several types of lane change are modeled under CSIM:
  1. from Local to Express, a change made when a CAV does not have a planned exit at the next node;
  2. from Express to Local, a change made when a CAV has a planned exit at the next node;
  3. from an entry lane to the Local Lane to join traffic; and
  4. from a Local Lane to an exit lane at an exit node.
Before changing lanes, a CAV senses whether it is safe to do so. To complete the lane change, the CAV adjusts velocity and position to preserve safe spacing in the new lane.
Forks and joins
Lane changes are envisioned to support forks and joins. As shown at right (above), to execute a fork, a CAV must be in the correct lane to take the path of choice, to the left or right. As shown at right (below), to execute a join where two lanes merge into one, a CAV must make a safe change when space in the slow (right) lane is available.

*A box is a construct used by CSIM to indicate a moving space that can contain a CAV. A CAV may only change lanes when there is an empty box in the adjacent (target) lane.


The lane change requirement is one that could be delayed as part of a transition strategy. While vehicles are still only partially-automated (e.g., dual-mode), lane changes could be left to drivers to execute. See also transition.






Example Lane-Change Protocol exploration
Assuming a 2-lane CAVWAY where node N(i) monitors traffic toward N(i+1), CAVs traveling in the CAVWAY at node N(i) fall into two groups: group A, consisting of CAVs exiting at node N(i+1); and group B, consisting of CAVs exiting beyond node N(i+1). It is then possible to envision four scenarios at node N(i).
  • Scenario 1: traffic is relatively light such that neither group A, which requires access to the right lane, nor group B, which can use the left lane, will exceed the capacity of the respective lanes.
  • Scenario 2: traffic is heavy such that there is at least enough demand from group A to fill the right lane to capacity and at least enough demand from group B to fill the left lane.
  • Scenario 3: demand from group A is heavy and demand from group B is light.
  • Scenario 4: demand from group A is light and demand from group B is heavy.
The following protocol addresses all four scenarios for the segment N(i) to N(i+1): CAVs exiting at node N(i+j) or earlier (where j > 0), are assigned to the right lane; CAVs exiting beyond N(i+j) are assigned to the left lane; and lanes are loaded to capacity or demand, whichever is less. This constraint applies: the right lane must always accommodate all CAVs exiting at node N(i+1).
For scenarios 1, 2, and 3, the value of j may be set to 1; in other words, CAVs exiting at node N(i+1) are assigned to the right lane while all others are assigned to the left lane. Node N(i) would need to control the number of CAVs entering the corridor such that neither lane in the segment was loaded beyond capacity. Under scenario 4, where more CAVs are traveling long distances, j could be set to a value greater than 1 to offload traffic from the left lane and take advantage of available space in the right lane. This situation would need to be monitored to assure that the value of j was always low enough to allow all exiting vehicles access to the right lane.
While more efficient approaches may yet be discovered, this exercise shows that it is possible to accommodate necessary CAV exits and to maintain (relatively) efficient use of the entire CAVWAY .
The protocol was verified using CSIM.