Semi-Autonomous Architectures for Safety and Dependence Mitigation

Autonomous control can aid an operator who is unable to safely manage a system. However, the status quo of automation has resulted in an untenable state. During a well understood mode of operation, automation can perform exquisitely, which generally leads to excessive dependence by the operator. Unfortunately existing automated procedures are unable to operate in arbitrary circumstances which can lead to catastrophe when an operator gets lazy. Our current research effort to combat this problem is twofold. First, we are focused on expanding the power of automation by improving the speed of optimization techniques and expanding their applicability. Second, we are investigating techniques for automatic intervention that mitigate operator dependence.


  1. R. Vasudevan, “Numerical Advection of Probability Densities for High Dimensional Systems,” in SIAM Conference on Applications of Dynamical Systems, 2015. 
  2. V. Shia, Y. Gao, R. Vasudevan, K. Campbell, T. Lin, F. Borrelli and R. Bajcsy, “Semiautonomous Vehicular Control Using Driver Modeling,” IEEE Transactions on Intelligent Transportation Systems, vol. 15, no. 6, pp. 2696-2709, November 2014. [url]
  3. V. Shia, R. Vasudevan, R. Bajcsy and R. Tedrake, “Convex Computation of the Reachable Set for Controlled Polynomial Hybrid Systems,” in IEEE Conference on Decision and Control, 2014. [pdf]
  4. R. Vasudevan, V. Shia, Y. Gao, R. Cervera-Navarro, R. Bajcsy and F. Borrelli, “Safe semi-autonomous control with enhanced driver modeling,” in American Control Conference, pp. 2896-2903, 2012. [pdf]
  5. A. Majumdar, R. Vasudevan, M. M. Tobenkin and R. Tedrake, “Convex Optimization of Nonlinear Feedback Controllers via Occupation Measures,” International Journal of Robotics Research, vol. 33, pp. 1209-1230, August 2014. [url]