A new method to enable vehicle mode detection in a cooperative environment in the absence of direct radio communication is proposed. The behavior or trajectory of a vehicle is described using a finite number of operating modes. In each mode the vehicle's trajectory is defined by a nominal desired reference and a perturbation signature based on Gold codes. To accomplish mode detection, a locally most powerful detector is derived, which can uniquely be evaluated using the Kalman Filter innovations. The performance of the locally most powerful detector is compared with a suboptimal, yet less computationally intensive alternative, via Monte Carlo simulations over the system parameter space. The effects of nonlinearities on detection performance are also investigated.