Algorithms for Cooperative and Connected Driving in the Presence of only few such Intelligent Vehicles

Negotiating and executing cooperative driving maneuvers with several fully automated vehicles provides considerable potential, for example, in terms of accident avoidance and emission reduction. However, these advantages will only be achieved once fully automated, cooperative vehicles have become introduced into traffic to a significant number of such vehicles. At the beginning though, these vehicles will not be able to find any cooperation partners and thus the cooperative function is not usable. Therefore, the question arises concerning the customer’s incentive to buy such a system for a possibly not inconsiderable surcharge.

Evolutionary Introduction of Cooperative Driving Functions

The project “iFORESEE – Introductory Concepts for Connected and Cooperative Cars“ of the “Profilregion Mobilitätssysteme Karlsruhe“ (“Profilregion: High Performance Center for Mobility Research”) addresses this question in an interdisciplinary approach of computer scientists, traffic engineers, psychologists and economic sociologists aiming for an evolutionary introduction of cooperative driving functions.

Essentially, this means:

  • The development of new cooperative driving functions that require low levels of automation (SAE level – from pure assistance functions, to the co-utilisation of, for example, lane keeping assistant or ACC, to automated driving functions of future highway pilots).
  • Cooperative driving functions that already offer an added value for its drivers even with only few such vehicles on the road
  • Driving functions that extend existing assistance functions to avoid high additional costs for additional sensor technology
  • Upwards-compatible driving functions to serve as initial cooperation partners for fully automated, cooperative vehicles and thus facilitate their market launch

As part of the “Karlsruhe Mobility Summit 2021“, two of the 13 developed driving functions were presented and simulated with OCTANE: the bottleneck-assistant and the rendezvous-assistant.


The Bottleneck-Assistant is a cooperative function that can be implemented in highly automated vehicles, but also as a pure assistance function without any automated vehicle driving control. The aim is to optimize traffic flow at bottlenecks without traffic lights by ensuring that vehicles equipped with the Bottleneck-Assistant give each other the right of way. Due to the special design of the algorithm including several of such intelligent vehicles, a harmonization of both driving directions is achieved without long waiting times to increase acceptance.

Video of the Bottleneck-Assistant  

The video demonstrates the optimized decision making of connected vehicles in low vehicle density.


In contrast to the Bottleneck-Assistant, the Rendezvous-Assistant targets vehicles with higher levels of automation that are already capable of fully-automated driving on freeways or expressways. One of the challenges here is changing between freeways at interchanges: It is to be expected that first highway pilots will require the driver to take over for such a maneuver, since fully automated merging has a high risk and a cooperative merging maneuver would most likely fail due to not finding a cooperation partner, as there will not be enough such intelligent vehicles in traffic at the beginning.

The Rendezvous-Assistant aims to close this gap by establishing communication with a suitable partner on the destination highway several kilometers before the intersection in order to enable the cooperative merging maneuver at the ramp. The algorithm synchronizes the arrival times of the two vehicles, taking into account the overall traffic flow and imposing speed requests onto the existing manufacturer-side speed assistance system. Thus, for example, a successful cooperation of a highway pilot aiming to change freeway with a vehicle with pure ACC functionality with the target speed coming from the Rendezvous-Assistent is possible.

Video of the Rendezvous-Assistant

The video demonstrates a traffic situation that can be imagined for the use of the Rendezvous Assistant, in which the meeting of two networked cars is calculated and controlled.

Generation of Synthetic Sensor Data and Ground Truth

Additionally, alongside with the two cooperative driving functions OCTANE was presented at the Karlsruhe Mobility Summit 2021 as a contribution to the “digitalization of testing” with its possibility of synthetic sensor data and ground truth generation.

Video of the Generation of Synthetic Sensor Data and Ground Truth

The video shows how OCTANE collects information about the environment and generates sensor data.