With the advent of 5G enabled Tactile Internet, wireless teleoperation systems with haptic feedback have attracted significant attentions in the communication society. One of the main challenges in this context is how to gain the maximum overall Quality of Experience (QoE) for teleoperation sessions sharing the same communication network. In the presence of communication delay, these teleoperation sessions use different control schemes to stabilize the remote interaction. These control schemes come with different trade offs concerning the required communication resources and the resulting QoE. Furthermore, their performance strongly depends on the actual network delay. In this paper, we focus on two typical control schemes: (1) teleoperation with the time-domain passivity approach (TDPA); (2) model-mediated teleoperation (MMT). We first study the human operator's QoE characteristic with respect to different control schemes and different round-trip delay settings. Based on this QoE-delay model, we propose a new uplink scheduling algorithm for haptic communication over 5G networks aiming at maximizing the total QoE of teleoperation sessions sharing the same radio resources. Simulation results confirm the efficiency of the nronosed scheduler.