Sediment traps are used for the protection of urban settlements at rivers in mountainous regions. These structures aim at the retention of sediment in the case of hazardous floods, but existing sediment traps tend to retain sediment also when the discharge is not hazardous to the downstream urban regions. This excessive retention of sediment causes an interruption of the river continuum that may lead to channel incision and the morphological depletion of downstream reaches. Another problem is the remobilization of formerly deposited sediment during a flood, which is addressed here in terms of the unwanted flushing of sediment traps. This research project aims at the development of sediment traps which are permeable up to a certain flood, but not susceptible to unwanted sediment flushing. Typical sediment traps consist of a retention area upstream of a barrier or check dam equipped with openings. The barrier can trigger the retention of sediment in the deposition area either by hydraulic control or by mechanical control. The hydraulic control leading to deposition is achieved by check dams with one or more openings constricting the flow vertically and/or laterally. Improved formulae for the estimation of the discharge capacity of such constrictions have been experimentally obtained for rough, turbulent upstream flow conditions with bed load under varying channel slopes. The constriction-induced head loss and reduction in the bed load transport capacity based on the bed shear stress are analyzed as a function of the upstream flow depth and discharge. The experiments show that the flushing of upstream sediment deposits may occur at open-crested slit check dams or close-crested slot check dams, but only when the latter are overtopped. The mechanical control leading to sediment retention is achieved by screens with vertical bars. The horizontal space between the bars corresponds to the characteristic grain size of traveling bed load. The required bottom clearance under such screens was optimized here in view of the possibility of bed load transfer for small (flood) discharges on the one hand, and the ensured clogging of the screen for high (flood) discharges on the other hand. This optimum bottom clearance height was found to be 1.75 times the characteristic grain size that is transported during floods. Once the bar screen was clogged, the unwanted sediment flushing could not occur anymore. However, the clogging depends on the estimation of the characteristic grain size. The experimental study shows that the combination of mechanical and hydraulic control structures provides a reliably working solution for permeable sediment traps. Smaller bed load-laden discharges can pass unhindered through such combined barriers. For higher discharges, the hydraulic control causes backwater which reduces the influence of the characteristic grain size on the clogging of the bar screen. Moreover, the bar screen prevents unwanted sediment flushing through the hydraulic control. The implementation of a guiding channel across the retention area is introduced and was experimentally verified as being a pertinent structural tool for improving the eco-morphological flow continuum. Finally, the design of a permeable sediment trap is described based on an optimal interaction between a guiding channel and a barrier combining the mechanical control by a bar screen and hydraulic control by a slot check dam.