The thesis presents methods for controlling and planning distributed energy resources (DERs) in active distribution networks (ADNs). It deals with three main challenges: (i) developing and experimentally validating grid-aware real-time control frameworks, (ii) estimating power grid's models from distributed measurements for measurement-based control schemes, and (iii) wide-scale planning of DERs in ADNs. The first part proposes real-time grid-aware control frameworks for ADNs hosting heterogeneous, controllable, and stochastic DERs. It presents a control and scheduling framework that tracks a pre-defined power profile (dispatch plan) at the ADN's grid connection point (GCP) while ensuring that the grid states (i.e., the nodal voltages and lines/transformer power/current flows) remain within the prescribed limits. A linear and accurate OPF-based real-time control scheme relying on power-flow sensitivity coefficients is proposed to obtain a tractable and computationally efficient formulation. The control scheme is experimentally validated on a real-scale 200kVA/0.4kV low-voltage microgrid hosted at the EPFL Distributed Electrical Systems Laboratory, representing a replica of the CIGRE benchmark system. Then, the thesis considers a case when the installed controllable resource is insufficient to cover the uncertainties caused by the stochastic injections. It proposes a two-layer MPC where an upper layer MPC, running at a slower timescale, optimizes battery SOC trajectories while minimizing the tracking error considering the forecast of the stochastic demand and generation for the whole day. And, a lower layer MPC, running at a faster timescale, takes battery SOC (from upper layer) trajectories as constraints while achieving high-resolution dispatch tracking. The control scheme uses the AR-OPF, an exact convex relaxation of AC-OPF, to model the grid constraints. The control scheme is experimentally validated using a 1.5MVA/2.5MWh BESS connected to an actual 24-node medium-voltage ADN hosting 3.2MWp photovoltaic (PV)- and 3.4MVA hydro-generations, and 2.8MW base demand. Thesis second part presents model-less or measurement-based control schemes where the network models are inferred from online measurements and then used in the control frameworks. In this respect, two methods are proposed. The first, referred to as the indirect method, estimates ADN's compound admittance matrix, and then computes the power-flow sensitivity coefficients. Second, a direct method to estimate the sensitivity coefficients using the nodal voltage and power measurements is presented. The direct and indirect methods are compared concerning estimation accuracy and variance. Thesis final part presents a planning tool to analyze the impact of large-scale integration of stochastic renewable resources on the planning of countrywide ADNs. In this respect, it presents a tool to generate countrywide synthetic power distribution networks using publicly available geo-referenced data. Then, it assesses the PV hosting capacity of distribution networks by formulating a grid-aware planning problem. An optimal planning problem of BESSs in ADNs is proposed to increase its hosting capacity. The OPF models developed previously are used to account for the grid constraints. A scheme for the cost-optimal countrywide deployment of PV and BESS units considering the MV power distribution infrastructure's limitations is proposed. The method is applied to Switzerland as a case study.