Statistical estimation of aperiodic response times when scheduled on top of static timelines

Safety-critical embedded control systems must support both periodic and stochastic functions on common hard-ware. Scheduling techniques that produce a timeline with predefined blocks for periodic functions are used to guaran-tee deadlines for critical closed loop periodic control func-tions (i.e. sensor read, control, actuator write). Aperiodic processing occurs in periodic timeline gaps. We present aperiodic response time estimation data using a new au-tomated binning technique based on a fluid flow analy-sis coupled with observations on conditional probabilities for response time values. Prior simulation results showed good response estimates under a broad range of conditions for a first-in first-out (fifo) aperiodic server. This paper presents new response time estimates, collected from a real